_ZN7roaring7RoaringC2EmPKj: 85| 12.4k| Roaring(size_t n, const uint32_t *data) : Roaring() { 86| 12.4k| api::roaring_bitmap_add_many(&roaring, n, data); 87| 12.4k| } _ZN7roaring7RoaringC2Ev: 76| 18.7k| Roaring() : roaring{} { 77| | // The empty constructor roaring{} silences warnings from pedantic 78| | // static analyzers. 79| 18.7k| api::roaring_bitmap_init_cleared(&roaring); 80| 18.7k| } _ZN7roaring7RoaringD2Ev: 913| 56.2k| ~Roaring() { 914| 56.2k| if (!(roaring.high_low_container.flags & ROARING_FLAG_FROZEN)) { ------------------ | | 47| 56.2k|#define ROARING_FLAG_FROZEN UINT8_C(0x2) ------------------ | Branch (914:13): [True: 56.2k, False: 0] ------------------ 915| 56.2k| api::roaring_bitmap_clear(&roaring); 916| 56.2k| } else { 917| | // The roaring member variable copies the `roaring_bitmap_t` and 918| | // nested `roaring_array_t` structures by value and is freed in the 919| | // constructor, however the underlying memory arena used for the 920| | // container data is not freed with it. Here we derive the arena 921| | // pointer from the second arena allocation in 922| | // `roaring_bitmap_frozen_view` and free it as well. 923| 0| roaring_bitmap_free( 924| 0| (roaring_bitmap_t *)((char *) 925| 0| roaring.high_low_container.containers - 926| 0| sizeof(roaring_bitmap_t))); 927| 0| } 928| 56.2k| } _ZN7roaring7Roaring11runOptimizeEv: 493| 12.4k| bool runOptimize() noexcept { 494| 12.4k| return api::roaring_bitmap_run_optimize(&roaring); 495| 12.4k| } _ZN7roaring7Roaring11shrinkToFitEv: 501| 6.24k| size_t shrinkToFit() noexcept { 502| 6.24k| return api::roaring_bitmap_shrink_to_fit(&roaring); 503| 6.24k| } _ZN7roaring7Roaring3addEj: 198| 6.24k| void add(uint32_t x) noexcept { api::roaring_bitmap_add(&roaring, x); } _ZN7roaring7Roaring10addCheckedEj: 205| 6.24k| bool addChecked(uint32_t x) noexcept { 206| 6.24k| return api::roaring_bitmap_add_checked(&roaring, x); 207| 6.24k| } _ZN7roaring7Roaring8addRangeEmm: 212| 6.24k| void addRange(const uint64_t min, const uint64_t max) noexcept { 213| 6.24k| return api::roaring_bitmap_add_range(&roaring, min, max); 214| 6.24k| } _ZN7roaring7Roaring7addManyEmPKj: 226| 6.24k| void addMany(size_t n_args, const uint32_t *vals) noexcept { 227| 6.24k| api::roaring_bitmap_add_many(&roaring, n_args, vals); 228| 6.24k| } _ZN7roaring7Roaring6removeEj: 258| 6.24k| void remove(uint32_t x) noexcept { 259| 6.24k| api::roaring_bitmap_remove(&roaring, x); 260| 6.24k| } _ZN7roaring7Roaring13removeCheckedEj: 267| 6.24k| bool removeChecked(uint32_t x) noexcept { 268| 6.24k| return api::roaring_bitmap_remove_checked(&roaring, x); 269| 6.24k| } _ZN7roaring7Roaring11removeRangeEmm: 274| 6.24k| void removeRange(uint64_t min, uint64_t max) noexcept { 275| 6.24k| return api::roaring_bitmap_remove_range(&roaring, min, max); 276| 6.24k| } _ZN7roaring7Roaring17removeRangeClosedEjj: 281| 6.24k| void removeRangeClosed(uint32_t min, uint32_t max) noexcept { 282| 6.24k| return api::roaring_bitmap_remove_range_closed(&roaring, min, max); 283| 6.24k| } _ZNK7roaring7Roaring7maximumEv: 315| 6.24k| uint32_t maximum() const noexcept { 316| 6.24k| return api::roaring_bitmap_maximum(&roaring); 317| 6.24k| } _ZNK7roaring7Roaring7minimumEv: 322| 6.24k| uint32_t minimum() const noexcept { 323| 6.24k| return api::roaring_bitmap_minimum(&roaring); 324| 6.24k| } _ZNK7roaring7Roaring8containsEj: 329| 6.24k| bool contains(uint32_t x) const noexcept { 330| 6.24k| return api::roaring_bitmap_contains(&roaring, x); 331| 6.24k| } _ZNK7roaring7Roaring13containsRangeEmm: 336| 6.24k| bool containsRange(const uint64_t x, const uint64_t y) const noexcept { 337| 6.24k| return api::roaring_bitmap_contains_range(&roaring, x, y); 338| 6.24k| } _ZNK7roaring7Roaring6selectEjPj: 526| 6.24k| bool select(uint32_t rnk, uint32_t *element) const noexcept { 527| 6.24k| return api::roaring_bitmap_select(&roaring, rnk, element); 528| 6.24k| } _ZNK7roaring7Roaring9intersectERKS0_: 540| 6.24k| bool intersect(const Roaring &r) const noexcept { 541| 6.24k| return api::roaring_bitmap_intersect(&roaring, &r.roaring); 542| 6.24k| } _ZNK7roaring7Roaring13jaccard_indexERKS0_: 551| 6.24k| double jaccard_index(const Roaring &r) const noexcept { 552| 6.24k| return api::roaring_bitmap_jaccard_index(&roaring, &r.roaring); 553| 6.24k| } _ZNK7roaring7Roaring14or_cardinalityERKS0_: 558| 6.24k| uint64_t or_cardinality(const Roaring &r) const noexcept { 559| 6.24k| return api::roaring_bitmap_or_cardinality(&roaring, &r.roaring); 560| 6.24k| } _ZNK7roaring7Roaring18andnot_cardinalityERKS0_: 565| 6.24k| uint64_t andnot_cardinality(const Roaring &r) const noexcept { 566| 6.24k| return api::roaring_bitmap_andnot_cardinality(&roaring, &r.roaring); 567| 6.24k| } _ZNK7roaring7Roaring15xor_cardinalityERKS0_: 573| 6.24k| uint64_t xor_cardinality(const Roaring &r) const noexcept { 574| 6.24k| return api::roaring_bitmap_xor_cardinality(&roaring, &r.roaring); 575| 6.24k| } _ZNK7roaring7Roaring4rankEj: 585| 6.24k| uint64_t rank(uint32_t x) const noexcept { 586| 6.24k| return api::roaring_bitmap_rank(&roaring, x); 587| 6.24k| } _ZNK7roaring7Roaring14getSizeInBytesEb: 732| 12.4k| size_t getSizeInBytes(bool portable = true) const noexcept { 733| 12.4k| if (portable) { ------------------ | Branch (733:13): [True: 12.4k, False: 0] ------------------ 734| 12.4k| return api::roaring_bitmap_portable_size_in_bytes(&roaring); 735| 12.4k| } else { 736| 0| return api::roaring_bitmap_size_in_bytes(&roaring); 737| 0| } 738| 12.4k| } _ZNK7roaring7RoaringanERKS0_: 794| 6.24k| Roaring operator&(const Roaring &o) const { 795| 6.24k| roaring_bitmap_t *r = api::roaring_bitmap_and(&roaring, &o.roaring); 796| 6.24k| if (r == NULL) { ------------------ | Branch (796:13): [True: 0, False: 6.24k] ------------------ 797| 0| ROARING_TERMINATE("failed materialization in and"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 798| 0| } 799| 6.24k| return Roaring(r); 800| 6.24k| } _ZN7roaring7RoaringC2EPNS_3api16roaring_bitmap_sE: 102| 31.2k| explicit Roaring(roaring_bitmap_t *s) noexcept : roaring(*s) { 103| 31.2k| roaring_free(s); // deallocate the passed-in pointer 104| 31.2k| } _ZNK7roaring7RoaringmiERKS0_: 807| 6.24k| Roaring operator-(const Roaring &o) const { 808| 6.24k| roaring_bitmap_t *r = api::roaring_bitmap_andnot(&roaring, &o.roaring); 809| 6.24k| if (r == NULL) { ------------------ | Branch (809:13): [True: 0, False: 6.24k] ------------------ 810| 0| ROARING_TERMINATE("failed materialization in andnot"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 811| 0| } 812| 6.24k| return Roaring(r); 813| 6.24k| } _ZNK7roaring7RoaringorERKS0_: 820| 6.24k| Roaring operator|(const Roaring &o) const { 821| 6.24k| roaring_bitmap_t *r = api::roaring_bitmap_or(&roaring, &o.roaring); 822| 6.24k| if (r == NULL) { ------------------ | Branch (822:13): [True: 0, False: 6.24k] ------------------ 823| 0| ROARING_TERMINATE("failed materialization in or"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 824| 0| } 825| 6.24k| return Roaring(r); 826| 6.24k| } _ZNK7roaring7RoaringeoERKS0_: 833| 6.24k| Roaring operator^(const Roaring &o) const { 834| 6.24k| roaring_bitmap_t *r = api::roaring_bitmap_xor(&roaring, &o.roaring); 835| 6.24k| if (r == NULL) { ------------------ | Branch (835:13): [True: 0, False: 6.24k] ------------------ 836| 0| ROARING_TERMINATE("failed materialization in xor"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 837| 0| } 838| 6.24k| return Roaring(r); 839| 6.24k| } _ZN7roaring7RoaringoRERKS0_: 378| 6.24k| Roaring &operator|=(const Roaring &r) noexcept { 379| 6.24k| api::roaring_bitmap_or_inplace(&roaring, &r.roaring); 380| 6.24k| return *this; 381| 6.24k| } _ZN7roaring7RoaringaNERKS0_: 356| 6.24k| Roaring &operator&=(const Roaring &r) noexcept { 357| 6.24k| api::roaring_bitmap_and_inplace(&roaring, &r.roaring); 358| 6.24k| return *this; 359| 6.24k| } _ZN7roaring7RoaringmIERKS0_: 366| 6.24k| Roaring &operator-=(const Roaring &r) noexcept { 367| 6.24k| api::roaring_bitmap_andnot_inplace(&roaring, &r.roaring); 368| 6.24k| return *this; 369| 6.24k| } _ZN7roaring7RoaringeOERKS0_: 388| 6.24k| Roaring &operator^=(const Roaring &r) noexcept { 389| 6.24k| api::roaring_bitmap_xor_inplace(&roaring, &r.roaring); 390| 6.24k| return *this; 391| 6.24k| } _ZNK7roaring7RoaringeqERKS0_: 458| 12.4k| bool operator==(const Roaring &r) const noexcept { 459| 12.4k| return api::roaring_bitmap_equals(&roaring, &r.roaring); 460| 12.4k| } _ZNK7roaring7Roaring11cardinalityEv: 401| 6.24k| uint64_t cardinality() const noexcept { 402| 6.24k| return api::roaring_bitmap_get_cardinality(&roaring); 403| 6.24k| } _ZNK7roaring7Roaring7isEmptyEv: 408| 12.4k| bool isEmpty() const noexcept { 409| 12.4k| return api::roaring_bitmap_is_empty(&roaring); 410| 12.4k| } _ZNK7roaring7Roaring13toUint32ArrayEPj: 441| 6.24k| void toUint32Array(uint32_t *ans) const noexcept { 442| 6.24k| api::roaring_bitmap_to_uint32_array(&roaring, ans); 443| 6.24k| } _ZNK7roaring7Roaring8isSubsetERKS0_: 425| 6.24k| bool isSubset(const Roaring &r) const noexcept { 426| 6.24k| return api::roaring_bitmap_is_subset(&roaring, &r.roaring); 427| 6.24k| } _ZNK7roaring7Roaring14isStrictSubsetERKS0_: 432| 6.24k| bool isStrictSubset(const Roaring &r) const noexcept { 433| 6.24k| return api::roaring_bitmap_is_strict_subset(&roaring, &r.roaring); 434| 6.24k| } _ZN7roaring7Roaring4flipEmm: 466| 6.24k| void flip(uint64_t range_start, uint64_t range_end) noexcept { 467| 6.24k| api::roaring_bitmap_flip_inplace(&roaring, range_start, range_end); 468| 6.24k| } _ZN7roaring7Roaring10flipClosedEjj: 474| 6.24k| void flipClosed(uint32_t range_start, uint32_t range_end) noexcept { 475| 6.24k| api::roaring_bitmap_flip_inplace_closed(&roaring, range_start, 476| 6.24k| range_end); 477| 6.24k| } _ZN7roaring7Roaring20removeRunCompressionEv: 483| 6.24k| bool removeRunCompression() noexcept { 484| 6.24k| return api::roaring_bitmap_remove_run_compression(&roaring); 485| 6.24k| } _ZN7roaring7RoaringC2ERKS0_: 110| 6.24k| Roaring(const Roaring &r) : Roaring() { 111| 6.24k| if (!api::roaring_bitmap_overwrite(&roaring, &r.roaring)) { ------------------ | Branch (111:13): [True: 0, False: 6.24k] ------------------ 112| 0| ROARING_TERMINATE("failed roaring_bitmap_overwrite in constructor"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 113| 0| } 114| 6.24k| api::roaring_bitmap_set_copy_on_write( 115| 6.24k| &roaring, api::roaring_bitmap_get_copy_on_write(&r.roaring)); 116| 6.24k| } _ZN7roaring7RoaringC2EOS0_: 122| 6.24k| Roaring(Roaring &&r) noexcept : roaring(r.roaring) { 123| | // 124| | // !!! This clones the bits of the roaring structure to a new location 125| | // and then overwrites the old bits...assuming that this will still 126| | // work. There are scenarios where this could break; e.g. if some of 127| | // those bits were pointers into the structure memory itself. If such 128| | // things were possible, a roaring_bitmap_move() API would be needed. 129| | // 130| 6.24k| api::roaring_bitmap_init_cleared(&r.roaring); 131| 6.24k| } _ZN7roaring7RoaringaSERKS0_: 152| 6.24k| Roaring &operator=(const Roaring &r) { 153| 6.24k| if (!api::roaring_bitmap_overwrite(&roaring, &r.roaring)) { ------------------ | Branch (153:13): [True: 0, False: 6.24k] ------------------ 154| 0| ROARING_TERMINATE("failed memory alloc in assignment"); ------------------ | | 27| 0|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 155| 0| } 156| 6.24k| api::roaring_bitmap_set_copy_on_write( 157| 6.24k| &roaring, api::roaring_bitmap_get_copy_on_write(&r.roaring)); 158| 6.24k| return *this; 159| 6.24k| } _ZN7roaring7RoaringaSEOS0_: 165| 6.24k| Roaring &operator=(Roaring &&r) noexcept { 166| 6.24k| api::roaring_bitmap_clear(&roaring); // free this class's allocations 167| | 168| | // !!! See notes in the Move Constructor regarding roaring_bitmap_move() 169| | // 170| 6.24k| roaring = r.roaring; 171| 6.24k| api::roaring_bitmap_init_cleared(&r.roaring); 172| | 173| 6.24k| return *this; 174| 6.24k| } _ZN7roaring7Roaring8readSafeEPKcm: 706| 12.4k| static Roaring readSafe(const char *buf, size_t maxbytes) { 707| 12.4k| roaring_bitmap_t *r = 708| 12.4k| api::roaring_bitmap_portable_deserialize_safe(buf, maxbytes); 709| 12.4k| if (r == NULL) { ------------------ | Branch (709:13): [True: 6.23k, False: 6.25k] ------------------ 710| 6.23k| ROARING_TERMINATE("failed alloc while reading"); ------------------ | | 27| 6.23k|#define ROARING_TERMINATE(_s) throw std::runtime_error(_s) ------------------ 711| 6.23k| } 712| 6.25k| return Roaring(r); 713| 12.4k| } _ZNK7roaring7Roaring5writeEPcb: 649| 6.24k| size_t write(char *buf, bool portable = true) const noexcept { 650| 6.24k| if (portable) { ------------------ | Branch (650:13): [True: 6.24k, False: 0] ------------------ 651| 6.24k| return api::roaring_bitmap_portable_serialize(&roaring, buf); 652| 6.24k| } else { 653| 0| return api::roaring_bitmap_serialize(&roaring, buf); 654| 0| } 655| 6.24k| } _ZNK7roaring7Roaring8toStringEv: 856| 6.24k| std::string toString() const noexcept { 857| 6.24k| struct iter_data { 858| 6.24k| std::string str{}; // The empty constructor silences warnings from 859| | // pedantic static analyzers. 860| 6.24k| char first_char = '{'; 861| 6.24k| } outer_iter_data; 862| 6.24k| if (!isEmpty()) { ------------------ | Branch (862:13): [True: 6.23k, False: 14] ------------------ 863| 6.23k| iterate( 864| 6.23k| [](uint32_t value, void *inner_iter_data) -> bool { 865| 6.23k| ((iter_data *)inner_iter_data)->str += 866| 6.23k| ((iter_data *)inner_iter_data)->first_char; 867| 6.23k| ((iter_data *)inner_iter_data)->str += 868| 6.23k| std::to_string(value); 869| 6.23k| ((iter_data *)inner_iter_data)->first_char = ','; 870| 6.23k| return true; 871| 6.23k| }, 872| 6.23k| (void *)&outer_iter_data); 873| 6.23k| } else 874| 14| outer_iter_data.str = '{'; 875| 6.24k| outer_iter_data.str += '}'; 876| 6.24k| return outer_iter_data.str; 877| 6.24k| } _ZNK7roaring7Roaring7iterateEPFbjPvES1_: 514| 6.23k| void iterate(api::roaring_iterator iterator, void *ptr) const { 515| 6.23k| api::roaring_iterate(&roaring, iterator, ptr); 516| 6.23k| } _ZZNK7roaring7Roaring8toStringEvENKUljPvE_clEjS1_: 864| 3.62G| [](uint32_t value, void *inner_iter_data) -> bool { 865| 3.62G| ((iter_data *)inner_iter_data)->str += 866| 3.62G| ((iter_data *)inner_iter_data)->first_char; 867| 3.62G| ((iter_data *)inner_iter_data)->str += 868| 3.62G| std::to_string(value); 869| 3.62G| ((iter_data *)inner_iter_data)->first_char = ','; 870| 3.62G| return true; 871| 3.62G| }, _ZNK7roaring7Roaring5beginEv: 1073| 12.4k|inline RoaringSetBitBiDirectionalIterator Roaring::begin() const { 1074| 12.4k| return RoaringSetBitBiDirectionalIterator(*this); 1075| 12.4k|} _ZN7roaring34RoaringSetBitBiDirectionalIteratorC2ERKNS_7RoaringEb: 967| 12.4k| bool exhausted = false) { 968| 12.4k| if (exhausted) { ------------------ | Branch (968:13): [True: 1, False: 12.4k] ------------------ 969| 1| i.parent = &parent.roaring; 970| 1| i.container_index = INT32_MAX; 971| 1| i.has_value = false; 972| 1| i.current_value = UINT32_MAX; 973| 12.4k| } else { 974| 12.4k| api::roaring_iterator_init(&parent.roaring, &i); 975| 12.4k| } 976| 12.4k| } _ZNK7roaring34RoaringSetBitBiDirectionalIteratorneERKS0_: 1064| 335k| bool operator!=(const RoaringSetBitBiDirectionalIterator &o) const { 1065| 335k| return i.current_value != *o || i.has_value != o.i.has_value; ------------------ | Branch (1065:16): [True: 329k, False: 6.24k] | Branch (1065:41): [True: 0, False: 6.24k] ------------------ 1066| 335k| } _ZNK7roaring34RoaringSetBitBiDirectionalIteratordeEv: 981| 335k| value_type operator*() const { return i.current_value; } _ZNK7roaring7Roaring3endEv: 1077| 335k|inline RoaringSetBitBiDirectionalIterator &Roaring::end() const { 1078| 335k| static RoaringSetBitBiDirectionalIterator e(*this, true); 1079| 335k| return e; 1080| 335k|} _ZN7roaring34RoaringSetBitBiDirectionalIteratorppEi: 1012| 329k| type_of_iterator operator++(int) { // i++, must return orig. value 1013| 329k| RoaringSetBitBiDirectionalIterator orig(*this); 1014| 329k| api::roaring_uint32_iterator_advance(&i); 1015| 329k| return orig; 1016| 329k| } _ZN7roaring34RoaringSetBitBiDirectionalIterator13equalorlargerEj: 1027| 6.24k| CROARING_DEPRECATED void equalorlarger(uint32_t val) { 1028| 6.24k| api::roaring_uint32_iterator_move_equalorlarger(&i, val); 1029| 6.24k| } binarySearch: 52| 7.14M| uint16_t ikey) { 53| 7.14M| const int32_t gap = 16; 54| 7.14M| if (lenarray < gap) { ------------------ | Branch (54:9): [True: 2.75M, False: 4.38M] ------------------ 55| 4.24M| for (int32_t j = 0; j < lenarray; j++) { ------------------ | Branch (55:29): [True: 4.24M, False: 7.03k] ------------------ 56| 4.24M| if (array[j] >= ikey) { ------------------ | Branch (56:17): [True: 2.75M, False: 1.49M] ------------------ 57| 2.75M| return (array[j] == ikey) ? j : -(j + 1); ------------------ | Branch (57:24): [True: 2.60M, False: 147k] ------------------ 58| 2.75M| } 59| 4.24M| } 60| 7.03k| return -(lenarray + 1); 61| 2.75M| } 62| 4.38M| const int32_t num_blocks = lenarray / gap; 63| 4.38M| int32_t base = 0; 64| 4.38M| int32_t n = num_blocks; 65| 8.85M| while (n > 3) { ------------------ | Branch (65:12): [True: 4.46M, False: 4.38M] ------------------ 66| 4.46M| int32_t quarter = n >> 2; 67| | 68| 4.46M| int32_t k1 = array[(base + quarter + 1) * gap - 1]; 69| 4.46M| int32_t k2 = array[(base + 2 * quarter + 1) * gap - 1]; 70| 4.46M| int32_t k3 = array[(base + 3 * quarter + 1) * gap - 1]; 71| | 72| 4.46M| int32_t c1 = (k1 < ikey); 73| 4.46M| int32_t c2 = (k2 < ikey); 74| 4.46M| int32_t c3 = (k3 < ikey); 75| | 76| 4.46M| base += (c1 + c2 + c3) * quarter; 77| 4.46M| n -= 3 * quarter; 78| 4.46M| } 79| 8.12M| while (n > 1) { ------------------ | Branch (79:12): [True: 3.74M, False: 4.38M] ------------------ 80| 3.74M| int32_t half = n >> 1; 81| 3.74M| base = (array[(base + half + 1) * gap - 1] < ikey) ? base + half : base; ------------------ | Branch (81:16): [True: 529k, False: 3.21M] ------------------ 82| 3.74M| n -= half; 83| 3.74M| } 84| 4.38M| int32_t lo = (array[(base + 1) * gap - 1] < ikey) ? base + 1 : base; ------------------ | Branch (84:18): [True: 2.07M, False: 2.30M] ------------------ 85| | 86| 4.38M| if (lo < num_blocks) { ------------------ | Branch (86:9): [True: 4.06M, False: 317k] ------------------ 87| 4.06M| const int32_t start = lo * gap; 88| 4.06M|#if defined(CROARING_IS_X64) 89| | // SSE2: subs_epu16 yields zero where lane >= ikey. movemask of an 90| | // epi16 compare gives 2 bits per lane; ctz>>1 = lane index. Scan 91| | // the first 8 lanes first and exit early when they contain the 92| | // answer; otherwise the block-narrowing invariant guarantees the 93| | // second-half mask is non-zero. 94| 4.06M| __m128i needle = _mm_set1_epi16((short)ikey); 95| 4.06M| __m128i zero = _mm_setzero_si128(); 96| 4.06M| __m128i v0 = _mm_loadu_si128((const __m128i *)(array + start)); 97| 4.06M| __m128i ge0 = _mm_cmpeq_epi16(_mm_subs_epu16(needle, v0), zero); 98| 4.06M| unsigned m0 = (unsigned)_mm_movemask_epi8(ge0); 99| 4.06M| if (m0 != 0) { ------------------ | Branch (99:13): [True: 2.88M, False: 1.18M] ------------------ 100| 2.88M| int32_t j = start + (int32_t)(roaring_trailing_zeroes(m0) >> 1); 101| 2.88M| return (array[j] == ikey) ? j : -(j + 1); ------------------ | Branch (101:20): [True: 2.12M, False: 758k] ------------------ 102| 2.88M| } 103| 1.18M| __m128i v1 = _mm_loadu_si128((const __m128i *)(array + start + 8)); 104| 1.18M| __m128i ge1 = _mm_cmpeq_epi16(_mm_subs_epu16(needle, v1), zero); 105| 1.18M| unsigned m1 = (unsigned)_mm_movemask_epi8(ge1); 106| 1.18M| int32_t j = start + 8 + (int32_t)(roaring_trailing_zeroes(m1) >> 1); 107| 1.18M| return (array[j] == ikey) ? j : -(j + 1); ------------------ | Branch (107:16): [True: 506k, False: 673k] ------------------ 108| |#else 109| | const int32_t end = start + gap; 110| | for (int32_t j = start; j < end; j++) { 111| | if (array[j] >= ikey) { 112| | return (array[j] == ikey) ? j : -(j + 1); 113| | } 114| | } 115| | // Unreachable: the narrowing guarantees the last element of the 116| | // selected block is >= ikey. 117| | return -(end + 1); 118| |#endif 119| 4.06M| } 120| | 121| 1.82M| for (int32_t j = num_blocks * gap; j < lenarray; j++) { ------------------ | Branch (121:40): [True: 1.81M, False: 10.4k] ------------------ 122| 1.81M| if (array[j] >= ikey) { ------------------ | Branch (122:13): [True: 306k, False: 1.51M] ------------------ 123| 306k| return (array[j] == ikey) ? j : -(j + 1); ------------------ | Branch (123:20): [True: 171k, False: 134k] ------------------ 124| 306k| } 125| 1.81M| } 126| 10.4k| return -(lenarray + 1); 127| 317k|} roaring.c:advanceUntil: 137| 469| int32_t length, uint16_t min) { 138| 469| int32_t lower = pos + 1; 139| | 140| 469| if ((lower >= length) || (array[lower] >= min)) { ------------------ | Branch (140:9): [True: 469, False: 0] | Branch (140:30): [True: 0, False: 0] ------------------ 141| 469| return lower; 142| 469| } 143| | 144| 0| int32_t spansize = 1; 145| | 146| 0| while ((lower + spansize < length) && (array[lower + spansize] < min)) { ------------------ | Branch (146:12): [True: 0, False: 0] | Branch (146:43): [True: 0, False: 0] ------------------ 147| 0| spansize <<= 1; 148| 0| } 149| 0| int32_t upper = (lower + spansize < length) ? lower + spansize : length - 1; ------------------ | Branch (149:21): [True: 0, False: 0] ------------------ 150| | 151| 0| if (array[upper] == min) { ------------------ | Branch (151:9): [True: 0, False: 0] ------------------ 152| 0| return upper; 153| 0| } 154| 0| if (array[upper] < min) { ------------------ | Branch (154:9): [True: 0, False: 0] ------------------ 155| | // means 156| | // array 157| | // has no 158| | // item 159| | // >= min 160| | // pos = array.length; 161| 0| return length; 162| 0| } 163| | 164| | // we know that the next-smallest span was too small 165| 0| lower += (spansize >> 1); 166| | 167| 0| int32_t mid = 0; 168| 0| while (lower + 1 != upper) { ------------------ | Branch (168:12): [True: 0, False: 0] ------------------ 169| 0| mid = (lower + upper) >> 1; 170| 0| if (array[mid] == min) { ------------------ | Branch (170:13): [True: 0, False: 0] ------------------ 171| 0| return mid; 172| 0| } else if (array[mid] < min) { ------------------ | Branch (172:20): [True: 0, False: 0] ------------------ 173| 0| lower = mid; 174| 0| } else { 175| 0| upper = mid; 176| 0| } 177| 0| } 178| 0| return upper; 179| 0|} roaring.c:count_greater: 197| 8.91k| uint16_t ikey) { 198| 8.91k| if (lenarray == 0) return 0; ------------------ | Branch (198:9): [True: 0, False: 8.91k] ------------------ 199| 8.91k| int32_t pos = binarySearch(array, lenarray, ikey); 200| 8.91k| if (pos >= 0) { ------------------ | Branch (200:9): [True: 1.93k, False: 6.97k] ------------------ 201| 1.93k| return lenarray - (pos + 1); 202| 6.97k| } else { 203| 6.97k| return lenarray - (-pos - 1); 204| 6.97k| } 205| 8.91k|} roaring.c:count_less: 186| 15.2k| uint16_t ikey) { 187| 15.2k| if (lenarray == 0) return 0; ------------------ | Branch (187:9): [True: 3.00k, False: 12.2k] ------------------ 188| 12.2k| int32_t pos = binarySearch(array, lenarray, ikey); 189| 12.2k| return pos >= 0 ? pos : -(pos + 1); ------------------ | Branch (189:12): [True: 9.40k, False: 2.87k] ------------------ 190| 15.2k|} array_util.c:advanceUntil: 137| 19.7k| int32_t length, uint16_t min) { 138| 19.7k| int32_t lower = pos + 1; 139| | 140| 19.7k| if ((lower >= length) || (array[lower] >= min)) { ------------------ | Branch (140:9): [True: 13, False: 19.7k] | Branch (140:30): [True: 7.05k, False: 12.6k] ------------------ 141| 7.06k| return lower; 142| 7.06k| } 143| | 144| 12.6k| int32_t spansize = 1; 145| | 146| 54.2k| while ((lower + spansize < length) && (array[lower + spansize] < min)) { ------------------ | Branch (146:12): [True: 52.8k, False: 1.34k] | Branch (146:43): [True: 41.5k, False: 11.3k] ------------------ 147| 41.5k| spansize <<= 1; 148| 41.5k| } 149| 12.6k| int32_t upper = (lower + spansize < length) ? lower + spansize : length - 1; ------------------ | Branch (149:21): [True: 11.3k, False: 1.34k] ------------------ 150| | 151| 12.6k| if (array[upper] == min) { ------------------ | Branch (151:9): [True: 3.28k, False: 9.37k] ------------------ 152| 3.28k| return upper; 153| 3.28k| } 154| 9.37k| if (array[upper] < min) { ------------------ | Branch (154:9): [True: 233, False: 9.13k] ------------------ 155| | // means 156| | // array 157| | // has no 158| | // item 159| | // >= min 160| | // pos = array.length; 161| 233| return length; 162| 233| } 163| | 164| | // we know that the next-smallest span was too small 165| 9.13k| lower += (spansize >> 1); 166| | 167| 9.13k| int32_t mid = 0; 168| 28.7k| while (lower + 1 != upper) { ------------------ | Branch (168:12): [True: 24.5k, False: 4.22k] ------------------ 169| 24.5k| mid = (lower + upper) >> 1; 170| 24.5k| if (array[mid] == min) { ------------------ | Branch (170:13): [True: 4.91k, False: 19.5k] ------------------ 171| 4.91k| return mid; 172| 19.5k| } else if (array[mid] < min) { ------------------ | Branch (172:20): [True: 9.62k, False: 9.96k] ------------------ 173| 9.62k| lower = mid; 174| 9.96k| } else { 175| 9.96k| upper = mid; 176| 9.96k| } 177| 24.5k| } 178| 4.22k| return upper; 179| 9.13k|} mixed_intersection.c:advanceUntil: 137| 137k| int32_t length, uint16_t min) { 138| 137k| int32_t lower = pos + 1; 139| | 140| 137k| if ((lower >= length) || (array[lower] >= min)) { ------------------ | Branch (140:9): [True: 536, False: 137k] | Branch (140:30): [True: 77.4k, False: 59.9k] ------------------ 141| 78.0k| return lower; 142| 78.0k| } 143| | 144| 59.9k| int32_t spansize = 1; 145| | 146| 146k| while ((lower + spansize < length) && (array[lower + spansize] < min)) { ------------------ | Branch (146:12): [True: 145k, False: 1.32k] | Branch (146:43): [True: 86.8k, False: 58.6k] ------------------ 147| 86.8k| spansize <<= 1; 148| 86.8k| } 149| 59.9k| int32_t upper = (lower + spansize < length) ? lower + spansize : length - 1; ------------------ | Branch (149:21): [True: 58.6k, False: 1.32k] ------------------ 150| | 151| 59.9k| if (array[upper] == min) { ------------------ | Branch (151:9): [True: 29.0k, False: 30.9k] ------------------ 152| 29.0k| return upper; 153| 29.0k| } 154| 30.9k| if (array[upper] < min) { ------------------ | Branch (154:9): [True: 635, False: 30.3k] ------------------ 155| | // means 156| | // array 157| | // has no 158| | // item 159| | // >= min 160| | // pos = array.length; 161| 635| return length; 162| 635| } 163| | 164| | // we know that the next-smallest span was too small 165| 30.3k| lower += (spansize >> 1); 166| | 167| 30.3k| int32_t mid = 0; 168| 54.6k| while (lower + 1 != upper) { ------------------ | Branch (168:12): [True: 38.6k, False: 15.9k] ------------------ 169| 38.6k| mid = (lower + upper) >> 1; 170| 38.6k| if (array[mid] == min) { ------------------ | Branch (170:13): [True: 14.3k, False: 24.2k] ------------------ 171| 14.3k| return mid; 172| 24.2k| } else if (array[mid] < min) { ------------------ | Branch (172:20): [True: 8.99k, False: 15.2k] ------------------ 173| 8.99k| lower = mid; 174| 15.2k| } else { 175| 15.2k| upper = mid; 176| 15.2k| } 177| 38.6k| } 178| 15.9k| return upper; 179| 30.3k|} mixed_subset.c:advanceUntil: 137| 436| int32_t length, uint16_t min) { 138| 436| int32_t lower = pos + 1; 139| | 140| 436| if ((lower >= length) || (array[lower] >= min)) { ------------------ | Branch (140:9): [True: 0, False: 436] | Branch (140:30): [True: 218, False: 218] ------------------ 141| 218| return lower; 142| 218| } 143| | 144| 218| int32_t spansize = 1; 145| | 146| 1.03k| while ((lower + spansize < length) && (array[lower + spansize] < min)) { ------------------ | Branch (146:12): [True: 954, False: 82] | Branch (146:43): [True: 818, False: 136] ------------------ 147| 818| spansize <<= 1; 148| 818| } 149| 218| int32_t upper = (lower + spansize < length) ? lower + spansize : length - 1; ------------------ | Branch (149:21): [True: 136, False: 82] ------------------ 150| | 151| 218| if (array[upper] == min) { ------------------ | Branch (151:9): [True: 0, False: 218] ------------------ 152| 0| return upper; 153| 0| } 154| 218| if (array[upper] < min) { ------------------ | Branch (154:9): [True: 10, False: 208] ------------------ 155| | // means 156| | // array 157| | // has no 158| | // item 159| | // >= min 160| | // pos = array.length; 161| 10| return length; 162| 10| } 163| | 164| | // we know that the next-smallest span was too small 165| 208| lower += (spansize >> 1); 166| | 167| 208| int32_t mid = 0; 168| 746| while (lower + 1 != upper) { ------------------ | Branch (168:12): [True: 538, False: 208] ------------------ 169| 538| mid = (lower + upper) >> 1; 170| 538| if (array[mid] == min) { ------------------ | Branch (170:13): [True: 0, False: 538] ------------------ 171| 0| return mid; 172| 538| } else if (array[mid] < min) { ------------------ | Branch (172:20): [True: 264, False: 274] ------------------ 173| 264| lower = mid; 174| 274| } else { 175| 274| upper = mid; 176| 274| } 177| 538| } 178| 208| return upper; 179| 208|} mixed_andnot.c:advanceUntil: 137| 94.2k| int32_t length, uint16_t min) { 138| 94.2k| int32_t lower = pos + 1; 139| | 140| 94.2k| if ((lower >= length) || (array[lower] >= min)) { ------------------ | Branch (140:9): [True: 4.79k, False: 89.4k] | Branch (140:30): [True: 76.2k, False: 13.1k] ------------------ 141| 81.0k| return lower; 142| 81.0k| } 143| | 144| 13.1k| int32_t spansize = 1; 145| | 146| 27.0k| while ((lower + spansize < length) && (array[lower + spansize] < min)) { ------------------ | Branch (146:12): [True: 26.5k, False: 468] | Branch (146:43): [True: 13.8k, False: 12.6k] ------------------ 147| 13.8k| spansize <<= 1; 148| 13.8k| } 149| 13.1k| int32_t upper = (lower + spansize < length) ? lower + spansize : length - 1; ------------------ | Branch (149:21): [True: 12.6k, False: 468] ------------------ 150| | 151| 13.1k| if (array[upper] == min) { ------------------ | Branch (151:9): [True: 6.62k, False: 6.54k] ------------------ 152| 6.62k| return upper; 153| 6.62k| } 154| 6.54k| if (array[upper] < min) { ------------------ | Branch (154:9): [True: 432, False: 6.11k] ------------------ 155| | // means 156| | // array 157| | // has no 158| | // item 159| | // >= min 160| | // pos = array.length; 161| 432| return length; 162| 432| } 163| | 164| | // we know that the next-smallest span was too small 165| 6.11k| lower += (spansize >> 1); 166| | 167| 6.11k| int32_t mid = 0; 168| 9.49k| while (lower + 1 != upper) { ------------------ | Branch (168:12): [True: 5.73k, False: 3.75k] ------------------ 169| 5.73k| mid = (lower + upper) >> 1; 170| 5.73k| if (array[mid] == min) { ------------------ | Branch (170:13): [True: 2.35k, False: 3.37k] ------------------ 171| 2.35k| return mid; 172| 3.37k| } else if (array[mid] < min) { ------------------ | Branch (172:20): [True: 1.15k, False: 2.21k] ------------------ 173| 1.15k| lower = mid; 174| 2.21k| } else { 175| 2.21k| upper = mid; 176| 2.21k| } 177| 5.73k| } 178| 3.75k| return upper; 179| 6.11k|} roaring.c:bitset_lenrange_cardinality: 63| 1.16k| uint32_t lenminusone) { 64| 1.16k| uint32_t firstword = start / 64; 65| 1.16k| uint32_t endword = (start + lenminusone) / 64; 66| 1.16k| if (firstword == endword) { ------------------ | Branch (66:9): [True: 856, False: 309] ------------------ 67| 856| return roaring_hamming(words[firstword] & 68| 856| ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 69| 856| << (start % 64)); 70| 856| } 71| 309| int answer = 72| 309| roaring_hamming(words[firstword] & ((~UINT64_C(0)) << (start % 64))); 73| 151k| for (uint32_t i = firstword + 1; i < endword; i++) { ------------------ | Branch (73:38): [True: 151k, False: 309] ------------------ 74| 151k| answer += roaring_hamming(words[i]); 75| 151k| } 76| 309| answer += roaring_hamming(words[endword] & 77| | (~UINT64_C(0)) >> 78| 309| (((~start + 1) - lenminusone - 1) % 64)); 79| 309| return answer; 80| 1.16k|} roaring.c:bitset_reset_range: 147| 1.13k| uint32_t end) { 148| 1.13k| if (start == end) return; ------------------ | Branch (148:9): [True: 0, False: 1.13k] ------------------ 149| 1.13k| uint32_t firstword = start / 64; 150| 1.13k| uint32_t endword = (end - 1) / 64; 151| 1.13k| if (firstword == endword) { ------------------ | Branch (151:9): [True: 856, False: 278] ------------------ 152| 856| words[firstword] &= ~(((~UINT64_C(0)) << (start % 64)) & 153| 856| ((~UINT64_C(0)) >> ((~end + 1) % 64))); 154| 856| return; 155| 856| } 156| 278| words[firstword] &= ~((~UINT64_C(0)) << (start % 64)); 157| 124k| for (uint32_t i = firstword + 1; i < endword; i++) { ------------------ | Branch (157:38): [True: 123k, False: 278] ------------------ 158| 123k| words[i] = UINT64_C(0); 159| 123k| } 160| | words[endword] &= ~((~UINT64_C(0)) >> ((~end + 1) % 64)); 161| 278|} roaring.c:bitset_set_lenrange: 112| 1.44k| uint32_t lenminusone) { 113| 1.44k| uint32_t firstword = start / 64; 114| 1.44k| uint32_t endword = (start + lenminusone) / 64; 115| 1.44k| if (firstword == endword) { ------------------ | Branch (115:9): [True: 0, False: 1.44k] ------------------ 116| 0| words[firstword] |= ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 117| 0| << (start % 64); 118| 0| return; 119| 0| } 120| 1.44k| uint64_t temp = words[endword]; 121| 1.44k| words[firstword] |= (~UINT64_C(0)) << (start % 64); 122| 427k| for (uint32_t i = firstword + 1; i < endword; i += 2) ------------------ | Branch (122:38): [True: 426k, False: 1.44k] ------------------ 123| 426k| words[i] = words[i + 1] = ~UINT64_C(0); 124| 1.44k| words[endword] = 125| | temp | (~UINT64_C(0)) >> (((~start + 1) - lenminusone - 1) % 64); 126| 1.44k|} bitset.c:avx2_harley_seal_popcount256: 355| 3.22k| const uint64_t size) { 356| 3.22k| __m256i total = _mm256_setzero_si256(); 357| 3.22k| __m256i ones = _mm256_setzero_si256(); 358| 3.22k| __m256i twos = _mm256_setzero_si256(); 359| 3.22k| __m256i fours = _mm256_setzero_si256(); 360| 3.22k| __m256i eights = _mm256_setzero_si256(); 361| 3.22k| __m256i sixteens = _mm256_setzero_si256(); 362| 3.22k| __m256i twosA, twosB, foursA, foursB, eightsA, eightsB; 363| | 364| 3.22k| const uint64_t limit = size - size % 16; 365| 3.22k| uint64_t i = 0; 366| | 367| 54.7k| for (; i < limit; i += 16) { ------------------ | Branch (367:12): [True: 51.5k, False: 3.22k] ------------------ 368| 51.5k| CSA(&twosA, &ones, ones, _mm256_lddqu_si256(data + i), 369| 51.5k| _mm256_lddqu_si256(data + i + 1)); 370| 51.5k| CSA(&twosB, &ones, ones, _mm256_lddqu_si256(data + i + 2), 371| 51.5k| _mm256_lddqu_si256(data + i + 3)); 372| 51.5k| CSA(&foursA, &twos, twos, twosA, twosB); 373| 51.5k| CSA(&twosA, &ones, ones, _mm256_lddqu_si256(data + i + 4), 374| 51.5k| _mm256_lddqu_si256(data + i + 5)); 375| 51.5k| CSA(&twosB, &ones, ones, _mm256_lddqu_si256(data + i + 6), 376| 51.5k| _mm256_lddqu_si256(data + i + 7)); 377| 51.5k| CSA(&foursB, &twos, twos, twosA, twosB); 378| 51.5k| CSA(&eightsA, &fours, fours, foursA, foursB); 379| 51.5k| CSA(&twosA, &ones, ones, _mm256_lddqu_si256(data + i + 8), 380| 51.5k| _mm256_lddqu_si256(data + i + 9)); 381| 51.5k| CSA(&twosB, &ones, ones, _mm256_lddqu_si256(data + i + 10), 382| 51.5k| _mm256_lddqu_si256(data + i + 11)); 383| 51.5k| CSA(&foursA, &twos, twos, twosA, twosB); 384| 51.5k| CSA(&twosA, &ones, ones, _mm256_lddqu_si256(data + i + 12), 385| 51.5k| _mm256_lddqu_si256(data + i + 13)); 386| 51.5k| CSA(&twosB, &ones, ones, _mm256_lddqu_si256(data + i + 14), 387| 51.5k| _mm256_lddqu_si256(data + i + 15)); 388| 51.5k| CSA(&foursB, &twos, twos, twosA, twosB); 389| 51.5k| CSA(&eightsB, &fours, fours, foursA, foursB); 390| 51.5k| CSA(&sixteens, &eights, eights, eightsA, eightsB); 391| | 392| 51.5k| total = _mm256_add_epi64(total, popcount256(sixteens)); 393| 51.5k| } 394| | 395| 3.22k| total = _mm256_slli_epi64(total, 4); // * 16 396| 3.22k| total = _mm256_add_epi64( 397| 3.22k| total, _mm256_slli_epi64(popcount256(eights), 3)); // += 8 * ... 398| 3.22k| total = _mm256_add_epi64( 399| 3.22k| total, _mm256_slli_epi64(popcount256(fours), 2)); // += 4 * ... 400| 3.22k| total = _mm256_add_epi64( 401| 3.22k| total, _mm256_slli_epi64(popcount256(twos), 1)); // += 2 * ... 402| 3.22k| total = _mm256_add_epi64(total, popcount256(ones)); 403| 3.22k| for (; i < size; i++) ------------------ | Branch (403:12): [True: 0, False: 3.22k] ------------------ 404| 0| total = 405| 0| _mm256_add_epi64(total, popcount256(_mm256_lddqu_si256(data + i))); 406| | 407| 3.22k| return (uint64_t)(_mm256_extract_epi64(total, 0)) + 408| 3.22k| (uint64_t)(_mm256_extract_epi64(total, 1)) + 409| 3.22k| (uint64_t)(_mm256_extract_epi64(total, 2)) + 410| | (uint64_t)(_mm256_extract_epi64(total, 3)); 411| 3.22k|} bitset.c:CSA: 343| 1.46M| __m256i c) { 344| 1.46M| const __m256i u = _mm256_xor_si256(a, b); 345| 1.46M| *h = _mm256_or_si256(_mm256_and_si256(a, b), _mm256_and_si256(u, c)); 346| 1.46M| *l = _mm256_xor_si256(u, c); 347| 1.46M|} bitset.c:popcount256: 307| 121k|static inline __m256i popcount256(__m256i v) { 308| 121k| const __m256i lookuppos = _mm256_setr_epi8( 309| 121k| /* 0 */ 4 + 0, /* 1 */ 4 + 1, /* 2 */ 4 + 1, /* 3 */ 4 + 2, 310| 121k| /* 4 */ 4 + 1, /* 5 */ 4 + 2, /* 6 */ 4 + 2, /* 7 */ 4 + 3, 311| 121k| /* 8 */ 4 + 1, /* 9 */ 4 + 2, /* a */ 4 + 2, /* b */ 4 + 3, 312| 121k| /* c */ 4 + 2, /* d */ 4 + 3, /* e */ 4 + 3, /* f */ 4 + 4, 313| | 314| | /* 0 */ 4 + 0, /* 1 */ 4 + 1, /* 2 */ 4 + 1, /* 3 */ 4 + 2, 315| 121k| /* 4 */ 4 + 1, /* 5 */ 4 + 2, /* 6 */ 4 + 2, /* 7 */ 4 + 3, 316| 121k| /* 8 */ 4 + 1, /* 9 */ 4 + 2, /* a */ 4 + 2, /* b */ 4 + 3, 317| 121k| /* c */ 4 + 2, /* d */ 4 + 3, /* e */ 4 + 3, /* f */ 4 + 4); 318| 121k| const __m256i lookupneg = _mm256_setr_epi8( 319| 121k| /* 0 */ 4 - 0, /* 1 */ 4 - 1, /* 2 */ 4 - 1, /* 3 */ 4 - 2, 320| 121k| /* 4 */ 4 - 1, /* 5 */ 4 - 2, /* 6 */ 4 - 2, /* 7 */ 4 - 3, 321| 121k| /* 8 */ 4 - 1, /* 9 */ 4 - 2, /* a */ 4 - 2, /* b */ 4 - 3, 322| 121k| /* c */ 4 - 2, /* d */ 4 - 3, /* e */ 4 - 3, /* f */ 4 - 4, 323| | 324| | /* 0 */ 4 - 0, /* 1 */ 4 - 1, /* 2 */ 4 - 1, /* 3 */ 4 - 2, 325| 121k| /* 4 */ 4 - 1, /* 5 */ 4 - 2, /* 6 */ 4 - 2, /* 7 */ 4 - 3, 326| 121k| /* 8 */ 4 - 1, /* 9 */ 4 - 2, /* a */ 4 - 2, /* b */ 4 - 3, 327| 121k| /* c */ 4 - 2, /* d */ 4 - 3, /* e */ 4 - 3, /* f */ 4 - 4); 328| 121k| const __m256i low_mask = _mm256_set1_epi8(0x0f); 329| | 330| 121k| const __m256i lo = _mm256_and_si256(v, low_mask); 331| 121k| const __m256i hi = _mm256_and_si256(_mm256_srli_epi16(v, 4), low_mask); 332| 121k| const __m256i popcnt1 = _mm256_shuffle_epi8(lookuppos, lo); 333| 121k| const __m256i popcnt2 = _mm256_shuffle_epi8(lookupneg, hi); 334| 121k| return _mm256_sad_epu8(popcnt1, popcnt2); 335| 121k|} convert.c:bitset_set_lenrange: 112| 106k| uint32_t lenminusone) { 113| 106k| uint32_t firstword = start / 64; 114| 106k| uint32_t endword = (start + lenminusone) / 64; 115| 106k| if (firstword == endword) { ------------------ | Branch (115:9): [True: 39.3k, False: 67.6k] ------------------ 116| 39.3k| words[firstword] |= ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 117| 39.3k| << (start % 64); 118| 39.3k| return; 119| 39.3k| } 120| 67.6k| uint64_t temp = words[endword]; 121| 67.6k| words[firstword] |= (~UINT64_C(0)) << (start % 64); 122| 33.3M| for (uint32_t i = firstword + 1; i < endword; i += 2) ------------------ | Branch (122:38): [True: 33.2M, False: 67.6k] ------------------ 123| 33.2M| words[i] = words[i + 1] = ~UINT64_C(0); 124| 67.6k| words[endword] = 125| | temp | (~UINT64_C(0)) >> (((~start + 1) - lenminusone - 1) % 64); 126| 67.6k|} mixed_intersection.c:bitset_lenrange_cardinality: 63| 201k| uint32_t lenminusone) { 64| 201k| uint32_t firstword = start / 64; 65| 201k| uint32_t endword = (start + lenminusone) / 64; 66| 201k| if (firstword == endword) { ------------------ | Branch (66:9): [True: 195k, False: 5.87k] ------------------ 67| 195k| return roaring_hamming(words[firstword] & 68| 195k| ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 69| 195k| << (start % 64)); 70| 195k| } 71| 5.87k| int answer = 72| 5.87k| roaring_hamming(words[firstword] & ((~UINT64_C(0)) << (start % 64))); 73| 6.15k| for (uint32_t i = firstword + 1; i < endword; i++) { ------------------ | Branch (73:38): [True: 280, False: 5.87k] ------------------ 74| 280| answer += roaring_hamming(words[i]); 75| 280| } 76| 5.87k| answer += roaring_hamming(words[endword] & 77| | (~UINT64_C(0)) >> 78| 5.87k| (((~start + 1) - lenminusone - 1) % 64)); 79| 5.87k| return answer; 80| 201k|} mixed_intersection.c:bitset_lenrange_empty: 86| 3.74k| uint32_t lenminusone) { 87| 3.74k| uint32_t firstword = start / 64; 88| 3.74k| uint32_t endword = (start + lenminusone) / 64; 89| 3.74k| if (firstword == endword) { ------------------ | Branch (89:9): [True: 3.46k, False: 279] ------------------ 90| 3.46k| return (words[firstword] & ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 91| 3.46k| << (start % 64)) == 0; 92| 3.46k| } 93| 279| if (((words[firstword] & ((~UINT64_C(0)) << (start % 64)))) != 0) { ------------------ | Branch (93:9): [True: 82, False: 197] ------------------ 94| 82| return false; 95| 82| } 96| 209| for (uint32_t i = firstword + 1; i < endword; i++) { ------------------ | Branch (96:38): [True: 14, False: 195] ------------------ 97| 14| if (words[i] != 0) { ------------------ | Branch (97:13): [True: 2, False: 12] ------------------ 98| 2| return false; 99| 2| } 100| 14| } 101| 195| if ((words[endword] & ------------------ | Branch (101:9): [True: 103, False: 92] ------------------ 102| 195| (~UINT64_C(0)) >> (((~start + 1) - lenminusone - 1) % 64)) != 0) { 103| 103| return false; 104| 103| } 105| 92| return true; 106| 195|} mixed_union.c:bitset_set_lenrange: 112| 100k| uint32_t lenminusone) { 113| 100k| uint32_t firstword = start / 64; 114| 100k| uint32_t endword = (start + lenminusone) / 64; 115| 100k| if (firstword == endword) { ------------------ | Branch (115:9): [True: 97.8k, False: 2.93k] ------------------ 116| 97.8k| words[firstword] |= ((~UINT64_C(0)) >> ((63 - lenminusone) % 64)) 117| 97.8k| << (start % 64); 118| 97.8k| return; 119| 97.8k| } 120| 2.93k| uint64_t temp = words[endword]; 121| 2.93k| words[firstword] |= (~UINT64_C(0)) << (start % 64); 122| 3.04k| for (uint32_t i = firstword + 1; i < endword; i += 2) ------------------ | Branch (122:38): [True: 108, False: 2.93k] ------------------ 123| 108| words[i] = words[i + 1] = ~UINT64_C(0); 124| 2.93k| words[endword] = 125| | temp | (~UINT64_C(0)) >> (((~start + 1) - lenminusone - 1) % 64); 126| 2.93k|} mixed_negation.c:bitset_flip_range: 132| 1.42k| uint32_t end) { 133| 1.42k| if (start == end) return; ------------------ | Branch (133:9): [True: 0, False: 1.42k] ------------------ 134| 1.42k| uint32_t firstword = start / 64; 135| 1.42k| uint32_t endword = (end - 1) / 64; 136| 1.42k| words[firstword] ^= ~((~UINT64_C(0)) << (start % 64)); 137| 184k| for (uint32_t i = firstword; i < endword; i++) { ------------------ | Branch (137:34): [True: 182k, False: 1.42k] ------------------ 138| 182k| words[i] = ~words[i]; 139| 182k| } 140| | words[endword] ^= ((~UINT64_C(0)) >> ((~end + 1) % 64)); 141| 1.42k|} mixed_xor.c:bitset_flip_range: 132| 53.8k| uint32_t end) { 133| 53.8k| if (start == end) return; ------------------ | Branch (133:9): [True: 0, False: 53.8k] ------------------ 134| 53.8k| uint32_t firstword = start / 64; 135| 53.8k| uint32_t endword = (end - 1) / 64; 136| 53.8k| words[firstword] ^= ~((~UINT64_C(0)) << (start % 64)); 137| 120k| for (uint32_t i = firstword; i < endword; i++) { ------------------ | Branch (137:34): [True: 66.7k, False: 53.8k] ------------------ 138| 66.7k| words[i] = ~words[i]; 139| 66.7k| } 140| | words[endword] ^= ((~UINT64_C(0)) >> ((~end + 1) % 64)); 141| 53.8k|} array_container_contains: 343| 3.25k| uint16_t pos) { 344| | /** 345| | * SIMD Quad algorithm 346| | * Daniel Lemire, "You can beat the binary search," in Daniel Lemire's blog, 347| | * April 27, 2026, 348| | * https://lemire.me/blog/2026/04/27/you-can-beat-the-binary-search/. 349| | */ 350| 3.25k| const int32_t gap = 16; 351| 3.25k| const uint16_t *carr = arr->array; 352| 3.25k| int32_t cardinality = arr->cardinality; 353| 3.25k| if (cardinality < gap) { ------------------ | Branch (353:9): [True: 949, False: 2.31k] ------------------ 354| 1.06k| for (int32_t j = 0; j < cardinality; j++) { ------------------ | Branch (354:29): [True: 1.04k, False: 20] ------------------ 355| 1.04k| if (carr[j] >= pos) return carr[j] == pos; ------------------ | Branch (355:17): [True: 929, False: 114] ------------------ 356| 1.04k| } 357| 20| return false; 358| 949| } 359| 2.31k| int32_t num_blocks = cardinality / gap; 360| 2.31k| int32_t base = 0; 361| 2.31k| int32_t n = num_blocks; 362| 5.24k| while (n > 3) { ------------------ | Branch (362:12): [True: 2.93k, False: 2.31k] ------------------ 363| 2.93k| int32_t quarter = n >> 2; 364| | 365| 2.93k| int32_t k1 = carr[(base + quarter + 1) * gap - 1]; 366| 2.93k| int32_t k2 = carr[(base + 2 * quarter + 1) * gap - 1]; 367| 2.93k| int32_t k3 = carr[(base + 3 * quarter + 1) * gap - 1]; 368| | 369| 2.93k| int32_t c1 = (k1 < pos); 370| 2.93k| int32_t c2 = (k2 < pos); 371| 2.93k| int32_t c3 = (k3 < pos); 372| | 373| 2.93k| base += (c1 + c2 + c3) * quarter; 374| 2.93k| n -= 3 * quarter; 375| 2.93k| } 376| 4.11k| while (n > 1) { ------------------ | Branch (376:12): [True: 1.80k, False: 2.31k] ------------------ 377| 1.80k| int32_t half = n >> 1; 378| 1.80k| base = (carr[(base + half + 1) * gap - 1] < pos) ? base + half : base; ------------------ | Branch (378:16): [True: 83, False: 1.71k] ------------------ 379| 1.80k| n -= half; 380| 1.80k| } 381| 2.31k| int32_t lo = (carr[(base + 1) * gap - 1] < pos) ? base + 1 : base; ------------------ | Branch (381:18): [True: 104, False: 2.20k] ------------------ 382| | 383| 2.31k| if (lo < num_blocks) { ------------------ | Branch (383:9): [True: 2.23k, False: 77] ------------------ 384| 2.23k| const uint16_t *blk = carr + lo * gap; 385| |#ifdef CROARING_USENEON 386| | uint16x8_t needle = vdupq_n_u16(pos); 387| | uint16x8_t v0 = vld1q_u16(blk); 388| | uint16x8_t v1 = vld1q_u16(blk + 8); 389| | uint16x8_t hit = 390| | vorrq_u16(vceqq_u16(v0, needle), vceqq_u16(v1, needle)); 391| | return vmaxvq_u16(hit) != 0; 392| |#elif defined(CROARING_IS_X64) 393| | __m128i needle = _mm_set1_epi16((short)pos); 394| 2.23k| __m128i v0 = _mm_loadu_si128((const __m128i *)blk); 395| 2.23k| __m128i v1 = _mm_loadu_si128((const __m128i *)(blk + 8)); 396| 2.23k| __m128i hit = _mm_or_si128(_mm_cmpeq_epi16(v0, needle), 397| 2.23k| _mm_cmpeq_epi16(v1, needle)); 398| 2.23k| return _mm_movemask_epi8(hit) != 0; 399| |#else 400| | for (int32_t j = 0; j < gap; j++) { 401| | if (blk[j] >= pos) return blk[j] == pos; 402| | } 403| | return false; 404| |#endif 405| 2.23k| } 406| | 407| 495| for (int32_t j = num_blocks * gap; j < cardinality; j++) { ------------------ | Branch (407:40): [True: 443, False: 52] ------------------ 408| 443| uint16_t v = carr[j]; 409| 443| if (v >= pos) return (v == pos); ------------------ | Branch (409:13): [True: 25, False: 418] ------------------ 410| 443| } 411| 52| return false; 412| 77|} array_container_minimum: 445| 3.54k|inline uint16_t array_container_minimum(const array_container_t *arr) { 446| 3.54k| if (arr->cardinality == 0) return 0; ------------------ | Branch (446:9): [True: 0, False: 3.54k] ------------------ 447| 3.54k| return arr->array[0]; 448| 3.54k|} array_container_maximum: 451| 12.9k|inline uint16_t array_container_maximum(const array_container_t *arr) { 452| 12.9k| if (arr->cardinality == 0) return 0; ------------------ | Branch (452:9): [True: 0, False: 12.9k] ------------------ 453| 12.9k| return arr->array[arr->cardinality - 1]; 454| 12.9k|} array_container_rank: 457| 3.70k|inline int array_container_rank(const array_container_t *arr, uint16_t x) { 458| 3.70k| const int32_t idx = binarySearch(arr->array, arr->cardinality, x); 459| 3.70k| const bool is_present = idx >= 0; 460| 3.70k| if (is_present) { ------------------ | Branch (460:9): [True: 3.13k, False: 563] ------------------ 461| 3.13k| return idx + 1; 462| 3.13k| } else { 463| 563| return -idx - 1; 464| 563| } 465| 3.70k|} array_container_index_equalorlarger: 507| 4.37k| uint16_t x) { 508| 4.37k| const int32_t idx = binarySearch(arr->array, arr->cardinality, x); 509| 4.37k| const bool is_present = idx >= 0; 510| 4.37k| if (is_present) { ------------------ | Branch (510:9): [True: 4.02k, False: 348] ------------------ 511| 4.02k| return idx; 512| 4.02k| } else { 513| 348| int32_t candidate = -idx - 1; 514| 348| if (candidate < arr->cardinality) return candidate; ------------------ | Branch (514:13): [True: 348, False: 0] ------------------ 515| 0| return -1; 516| 348| } 517| 4.37k|} roaring.c:array_container_add_range_nvals: 527| 582| int32_t nvals_greater) { 528| 582| int32_t union_cardinality = nvals_less + (max - min + 1) + nvals_greater; 529| 582| if (union_cardinality > array->capacity) { ------------------ | Branch (529:9): [True: 541, False: 41] ------------------ 530| 541| array_container_grow(array, union_cardinality, true); 531| 541| } 532| 582| memmove(&(array->array[union_cardinality - nvals_greater]), 533| 582| &(array->array[array->cardinality - nvals_greater]), 534| 582| nvals_greater * sizeof(uint16_t)); 535| 1.40M| for (uint32_t i = 0; i <= max - min; i++) { ------------------ | Branch (535:26): [True: 1.40M, False: 582] ------------------ 536| 1.40M| array->array[nvals_less + i] = (uint16_t)(min + i); 537| 1.40M| } 538| 582| array->cardinality = union_cardinality; 539| 582|} roaring.c:array_container_remove_range: 557| 3.14k| uint32_t pos, uint32_t count) { 558| 3.14k| if (count != 0) { ------------------ | Branch (558:9): [True: 283, False: 2.86k] ------------------ 559| 283| memmove(&(array->array[pos]), &(array->array[pos + count]), 560| 283| (array->cardinality - pos - count) * sizeof(uint16_t)); 561| 283| array->cardinality -= count; 562| 283| } 563| 3.14k|} roaring.c:array_container_cardinality: 82| 69.4k|static inline int array_container_cardinality(const array_container_t *array) { 83| 69.4k| return array->cardinality; 84| 69.4k|} roaring.c:array_container_try_add: 293| 7.12M| int32_t max_cardinality) { 294| 7.12M| const int32_t cardinality = arr->cardinality; 295| | 296| | // best case, we can append. 297| 7.12M| if ((array_container_empty(arr) || arr->array[cardinality - 1] < value) && ------------------ | Branch (297:10): [True: 15.6k, False: 7.10M] | Branch (297:40): [True: 50.3k, False: 7.05M] ------------------ 298| 65.9k| cardinality < max_cardinality) { ------------------ | Branch (298:9): [True: 65.9k, False: 0] ------------------ 299| 65.9k| array_container_append(arr, value); 300| 65.9k| return 1; 301| 65.9k| } 302| | 303| 7.05M| const int32_t loc = binarySearch(arr->array, cardinality, value); 304| | 305| 7.05M| if (loc >= 0) { ------------------ | Branch (305:9): [True: 5.36M, False: 1.69M] ------------------ 306| 5.36M| return 0; 307| 5.36M| } else if (cardinality < max_cardinality) { ------------------ | Branch (307:16): [True: 1.69M, False: 120] ------------------ 308| 1.69M| if (array_container_full(arr)) { ------------------ | Branch (308:13): [True: 75.1k, False: 1.61M] ------------------ 309| 75.1k| array_container_grow(arr, arr->capacity + 1, true); 310| 75.1k| } 311| 1.69M| const int32_t insert_idx = -loc - 1; 312| 1.69M| memmove(arr->array + insert_idx + 1, arr->array + insert_idx, 313| 1.69M| (cardinality - insert_idx) * sizeof(uint16_t)); 314| 1.69M| arr->array[insert_idx] = value; 315| 1.69M| arr->cardinality++; 316| 1.69M| return 1; 317| 1.69M| } else { 318| 120| return -1; 319| 120| } 320| 7.05M|} roaring.c:array_container_empty: 100| 7.12M|static inline bool array_container_empty(const array_container_t *array) { 101| 7.12M| return array->cardinality == 0; 102| 7.12M|} roaring.c:array_container_append: 274| 65.9k| uint16_t pos) { 275| 65.9k| const int32_t capacity = arr->capacity; 276| | 277| 65.9k| if (array_container_full(arr)) { ------------------ | Branch (277:9): [True: 31.5k, False: 34.4k] ------------------ 278| 31.5k| array_container_grow(arr, capacity + 1, true); 279| 31.5k| } 280| | 281| 65.9k| arr->array[arr->cardinality++] = pos; 282| 65.9k|} roaring.c:array_container_full: 106| 1.75M|static inline bool array_container_full(const array_container_t *array) { 107| 1.75M| return array->cardinality == array->capacity; 108| 1.75M|} roaring.c:array_container_remove: 329| 6.04k| uint16_t pos) { 330| 6.04k| const int32_t idx = binarySearch(arr->array, arr->cardinality, pos); 331| 6.04k| const bool is_present = idx >= 0; 332| 6.04k| if (is_present) { ------------------ | Branch (332:9): [True: 3.03k, False: 3.01k] ------------------ 333| 3.03k| memmove(arr->array + idx, arr->array + idx + 1, 334| 3.03k| (arr->cardinality - idx - 1) * sizeof(uint16_t)); 335| 3.03k| arr->cardinality--; 336| 3.03k| } 337| | 338| 6.04k| return is_present; 339| 6.04k|} roaring.c:array_container_nonzero_cardinality: 87| 100k| const array_container_t *array) { 88| 100k| return array->cardinality > 0; 89| 100k|} roaring.c:array_container_equals: 230| 9.23k| const array_container_t *container2) { 231| 9.23k| if (container1->cardinality != container2->cardinality) { ------------------ | Branch (231:9): [True: 2.27k, False: 6.96k] ------------------ 232| 2.27k| return false; 233| 2.27k| } 234| 6.96k| return memequals(container1->array, container2->array, 235| 6.96k| container1->cardinality * 2); 236| 9.23k|} roaring.c:array_container_select: 252| 3.70k| uint32_t *element) { 253| 3.70k| int card = array_container_cardinality(container); 254| 3.70k| if (*start_rank + card <= rank) { ------------------ | Branch (254:9): [True: 349, False: 3.35k] ------------------ 255| 349| *start_rank += card; 256| 349| return false; 257| 3.35k| } else { 258| 3.35k| *element = container->array[rank - *start_rank]; 259| | return true; 260| 3.35k| } 261| 3.70k|} roaring.c:array_container_contains_range: 421| 83| uint32_t range_end) { 422| 83| const int32_t range_count = range_end - range_start; 423| 83| const uint16_t rs_included = (uint16_t)range_start; 424| 83| const uint16_t re_included = (uint16_t)(range_end - 1); 425| | 426| | // Empty range is always included 427| 83| if (range_count <= 0) { ------------------ | Branch (427:9): [True: 0, False: 83] ------------------ 428| 0| return true; 429| 0| } 430| 83| if (range_count > arr->cardinality) { ------------------ | Branch (430:9): [True: 34, False: 49] ------------------ 431| 34| return false; 432| 34| } 433| | 434| 49| const int32_t start = 435| 49| binarySearch(arr->array, arr->cardinality, rs_included); 436| | // If this sorted array contains all items in the range: 437| | // * the start item must be found 438| | // * the last item in range range_count must exist, and be the expected end 439| | // value 440| 49| return (start >= 0) && (arr->cardinality >= start + range_count) && ------------------ | Branch (440:12): [True: 39, False: 10] | Branch (440:28): [True: 35, False: 4] ------------------ 441| 35| (arr->array[start + range_count - 1] == re_included); ------------------ | Branch (441:12): [True: 10, False: 25] ------------------ 442| 83|} roaring_array.c:array_container_size_in_bytes: 221| 17.3k| const array_container_t *container) { 222| 17.3k| return container->cardinality * sizeof(uint16_t); 223| 17.3k|} roaring_array.c:array_container_cardinality: 82| 6.82k|static inline int array_container_cardinality(const array_container_t *array) { 83| 6.82k| return array->cardinality; 84| 6.82k|} array.c:array_container_size_in_bytes: 221| 13.6k| const array_container_t *container) { 222| 13.6k| return container->cardinality * sizeof(uint16_t); 223| 13.6k|} convert.c:array_container_cardinality: 82| 14.7k|static inline int array_container_cardinality(const array_container_t *array) { 83| 14.7k| return array->cardinality; 84| 14.7k|} convert.c:array_container_serialized_size_in_bytes: 173| 142k|static inline int32_t array_container_serialized_size_in_bytes(int32_t card) { 174| 142k| return card * sizeof(uint16_t); 175| 142k|} mixed_intersection.c:array_container_empty: 100| 24|static inline bool array_container_empty(const array_container_t *array) { 101| 24| return array->cardinality == 0; 102| 24|} mixed_andnot.c:array_container_cardinality: 82| 368|static inline int array_container_cardinality(const array_container_t *array) { 83| 368| return array->cardinality; 84| 368|} bitset_container_get: 197| 1.04M| uint16_t pos) { 198| 1.04M| const uint64_t word = bitset->words[pos >> 6]; 199| 1.04M| return (word >> (pos & 63)) & 1; 200| 1.04M|} bitset_container_contains: 236| 905k| uint16_t pos) { 237| 905k| return bitset_container_get(bitset, pos); 238| 905k|} roaring.c:bitset_container_cardinality: 252| 10.4k| const bitset_container_t *bitset) { 253| 10.4k| return bitset->cardinality; 254| 10.4k|} roaring.c:bitset_container_set: 151| 303k| uint16_t pos) { 152| 303k| const uint64_t old_word = bitset->words[pos >> 6]; 153| 303k| const int index = pos & 63; 154| | const uint64_t new_word = old_word | (UINT64_C(1) << index); 155| 303k| bitset->cardinality += (uint32_t)((old_word ^ new_word) >> index); 156| 303k| bitset->words[pos >> 6] = new_word; 157| 303k|} roaring.c:bitset_container_add: 172| 120| uint16_t pos) { 173| 120| const uint64_t old_word = bitset->words[pos >> 6]; 174| 120| const int index = pos & 63; 175| | const uint64_t new_word = old_word | (UINT64_C(1) << index); 176| 120| const uint64_t increment = (old_word ^ new_word) >> index; 177| 120| bitset->cardinality += (uint32_t)increment; 178| 120| bitset->words[pos >> 6] = new_word; 179| 120| return increment > 0; 180| 120|} roaring.c:bitset_container_remove: 185| 2.15k| uint16_t pos) { 186| 2.15k| const uint64_t old_word = bitset->words[pos >> 6]; 187| 2.15k| const int index = pos & 63; 188| | const uint64_t new_word = old_word & (~(UINT64_C(1) << index)); 189| 2.15k| const uint64_t increment = (old_word ^ new_word) >> index; 190| 2.15k| bitset->cardinality -= (uint32_t)increment; 191| 2.15k| bitset->words[pos >> 6] = new_word; 192| 2.15k| return increment > 0; 193| 2.15k|} roaring.c:bitset_container_const_nonzero_cardinality: 285| 6.54k| const bitset_container_t *bitset) { 286| 6.54k| return !bitset_container_empty(bitset); 287| 6.54k|} roaring.c:bitset_container_empty: 272| 6.54k|static inline bool bitset_container_empty(const bitset_container_t *bitset) { 273| 6.54k| if (bitset->cardinality == BITSET_UNKNOWN_CARDINALITY) { ------------------ | Branch (273:9): [True: 0, False: 6.54k] ------------------ 274| 0| for (int i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; i++) { ------------------ | Branch (274:25): [True: 0, False: 0] ------------------ 275| 0| if ((bitset->words[i]) != 0) return false; ------------------ | Branch (275:17): [True: 0, False: 0] ------------------ 276| 0| } 277| 0| return true; 278| 0| } 279| 6.54k| return bitset->cardinality == 0; 280| 6.54k|} roaring.c:bitset_container_get_range: 210| 219| uint32_t pos_end) { 211| 219| const uint32_t start = pos_start >> 6; 212| 219| const uint32_t end = pos_end >> 6; 213| | 214| 219| const uint64_t first = ~((1ULL << (pos_start & 0x3F)) - 1); 215| 219| const uint64_t last = (1ULL << (pos_end & 0x3F)) - 1; 216| | 217| 219| if (start == end) ------------------ | Branch (217:9): [True: 5, False: 214] ------------------ 218| 5| return ((bitset->words[end] & first & last) == (first & last)); 219| 214| if ((bitset->words[start] & first) != first) return false; ------------------ | Branch (219:9): [True: 79, False: 135] ------------------ 220| | 221| 135| if ((end < BITSET_CONTAINER_SIZE_IN_WORDS) && ------------------ | Branch (221:9): [True: 117, False: 18] ------------------ 222| 117| ((bitset->words[end] & last) != last)) { ------------------ | Branch (222:9): [True: 95, False: 22] ------------------ 223| 95| return false; 224| 95| } 225| | 226| 40| for (uint32_t i = start + 1; 227| 15.7k| (i < BITSET_CONTAINER_SIZE_IN_WORDS) && (i < end); ++i) { ------------------ | Branch (227:10): [True: 15.7k, False: 15] | Branch (227:50): [True: 15.7k, False: 20] ------------------ 228| 15.7k| if (bitset->words[i] != UINT64_C(0xFFFFFFFFFFFFFFFF)) return false; ------------------ | Branch (228:13): [True: 5, False: 15.7k] ------------------ 229| 15.7k| } 230| | 231| 35| return true; 232| 40|} roaring_array.c:bitset_container_size_in_bytes: 472| 133k| const bitset_container_t *container) { 473| 133k| (void)container; 474| 133k| return BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t); 475| 133k|} roaring_array.c:bitset_container_cardinality: 252| 66.9k| const bitset_container_t *bitset) { 253| 66.9k| return bitset->cardinality; 254| 66.9k|} bitset.c:bitset_container_size_in_bytes: 472| 133k| const bitset_container_t *container) { 473| 133k| (void)container; 474| 133k| return BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t); 475| 133k|} convert.c:bitset_container_set: 151| 2.28M| uint16_t pos) { 152| 2.28M| const uint64_t old_word = bitset->words[pos >> 6]; 153| 2.28M| const int index = pos & 63; 154| | const uint64_t new_word = old_word | (UINT64_C(1) << index); 155| 2.28M| bitset->cardinality += (uint32_t)((old_word ^ new_word) >> index); 156| 2.28M| bitset->words[pos >> 6] = new_word; 157| 2.28M|} convert.c:bitset_container_serialized_size_in_bytes: 431| 130k|static inline int32_t bitset_container_serialized_size_in_bytes(void) { 432| 130k| return BITSET_CONTAINER_SIZE_IN_WORDS * 8; 433| 130k|} container_iterator_next: 2457| 329k| uint16_t *value) { 2458| 329k| switch (typecode) { 2459| 0| case BITSET_CONTAINER_TYPE: { ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2459:9): [True: 0, False: 329k] ------------------ 2460| 0| const bitset_container_t *bc = const_CAST_bitset(c); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2461| 0| it->index++; 2462| | 2463| 0| uint32_t wordindex = it->index / 64; 2464| 0| if (wordindex >= BITSET_CONTAINER_SIZE_IN_WORDS) { ------------------ | Branch (2464:17): [True: 0, False: 0] ------------------ 2465| 0| return false; 2466| 0| } 2467| | 2468| 0| uint64_t word = 2469| 0| bc->words[wordindex] & (UINT64_MAX << (it->index % 64)); 2470| | // next part could be optimized/simplified 2471| 0| while (word == 0 && ------------------ | Branch (2471:20): [True: 0, False: 0] ------------------ 2472| 0| (wordindex + 1 < BITSET_CONTAINER_SIZE_IN_WORDS)) { ------------------ | Branch (2472:20): [True: 0, False: 0] ------------------ 2473| 0| wordindex++; 2474| 0| word = bc->words[wordindex]; 2475| 0| } 2476| 0| if (word != 0) { ------------------ | Branch (2476:17): [True: 0, False: 0] ------------------ 2477| 0| it->index = wordindex * 64 + roaring_trailing_zeroes(word); 2478| 0| *value = it->index; 2479| 0| return true; 2480| 0| } 2481| 0| return false; 2482| 0| } 2483| 283k| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 283k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2483:9): [True: 283k, False: 45.3k] ------------------ 2484| 283k| const array_container_t *ac = const_CAST_array(c); ------------------ | | 55| 283k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 283k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2485| 283k| it->index++; 2486| 283k| if (it->index < ac->cardinality) { ------------------ | Branch (2486:17): [True: 278k, False: 5.55k] ------------------ 2487| 278k| *value = ac->array[it->index]; 2488| 278k| return true; 2489| 278k| } 2490| 5.55k| return false; 2491| 283k| } 2492| 45.3k| case RUN_CONTAINER_TYPE: { ------------------ | | 50| 45.3k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2492:9): [True: 45.3k, False: 283k] ------------------ 2493| 45.3k| if (*value == UINT16_MAX) { // Avoid overflow to zero ------------------ | Branch (2493:17): [True: 0, False: 45.3k] ------------------ 2494| 0| return false; 2495| 0| } 2496| | 2497| 45.3k| const run_container_t *rc = const_CAST_run(c); ------------------ | | 78| 45.3k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 45.3k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2498| 45.3k| uint32_t limit = 2499| 45.3k| rc->runs[it->index].value + rc->runs[it->index].length; 2500| 45.3k| if (*value < limit) { ------------------ | Branch (2500:17): [True: 28.4k, False: 16.9k] ------------------ 2501| 28.4k| (*value)++; 2502| 28.4k| return true; 2503| 28.4k| } 2504| | 2505| 16.9k| it->index++; 2506| 16.9k| if (it->index < rc->n_runs) { ------------------ | Branch (2506:17): [True: 16.4k, False: 489] ------------------ 2507| 16.4k| *value = rc->runs[it->index].value; 2508| 16.4k| return true; 2509| 16.4k| } 2510| 489| return false; 2511| 16.9k| } 2512| 0| default: ------------------ | Branch (2512:9): [True: 0, False: 329k] ------------------ 2513| 0| assert(false); 2514| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2515| 0| return false; 2516| 329k| } 2517| 329k|} roaring.c:container_from_range: 317| 68.2k| uint32_t max, uint16_t step) { 318| 68.2k| if (step == 0) return NULL; // being paranoid ------------------ | Branch (318:9): [True: 0, False: 68.2k] ------------------ 319| 68.2k| if (step == 1) { ------------------ | Branch (319:9): [True: 68.2k, False: 0] ------------------ 320| 68.2k| return container_range_of_ones(min, max, type); 321| | // Note: the result is not always a run (need to check the cardinality) 322| | //*type = RUN_CONTAINER_TYPE; 323| | // return run_container_create_range(min, max); 324| 68.2k| } 325| 0| int size = (max - min + step - 1) / step; 326| 0| if (size <= DEFAULT_MAX_SIZE) { // array container ------------------ | Branch (326:9): [True: 0, False: 0] ------------------ 327| 0| *type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 328| 0| array_container_t *array = array_container_create_given_capacity(size); 329| 0| array_container_add_from_range(array, min, max, step); 330| 0| assert(array->cardinality == size); 331| 0| return array; 332| 0| } else { // bitset container 333| 0| *type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 334| 0| bitset_container_t *bitset = bitset_container_create(); 335| 0| bitset_container_add_from_range(bitset, min, max, step); 336| | assert(bitset->cardinality == size); 337| 0| return bitset; 338| 0| } 339| 0|} roaring.c:container_range_of_ones: 302| 85.1k| uint8_t *result_type) { 303| 85.1k| assert(range_end >= range_start); 304| 85.1k| uint64_t cardinality = range_end - range_start + 1; 305| 85.1k| if (cardinality <= 2) { ------------------ | Branch (305:9): [True: 126, False: 85.0k] ------------------ 306| 126| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 126|#define ARRAY_CONTAINER_TYPE 2 ------------------ 307| 126| return array_container_create_range(range_start, range_end); 308| 85.0k| } else { 309| 85.0k| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 85.0k|#define RUN_CONTAINER_TYPE 3 ------------------ 310| 85.0k| return run_container_create_range(range_start, range_end); 311| 85.0k| } 312| 85.1k|} roaring.c:container_add_range: 2282| 3.16k| uint8_t *result_type) { 2283| | // NB: when selecting new container type, we perform only inexpensive checks 2284| 3.16k| switch (type) { 2285| 0| case BITSET_CONTAINER_TYPE: { ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2285:9): [True: 0, False: 3.16k] ------------------ 2286| 0| bitset_container_t *bitset = CAST_bitset(c); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2287| | 2288| 0| int32_t union_cardinality = 0; 2289| 0| union_cardinality += bitset->cardinality; 2290| 0| union_cardinality += max - min + 1; 2291| 0| union_cardinality -= 2292| 0| bitset_lenrange_cardinality(bitset->words, min, max - min); 2293| | 2294| 0| if (union_cardinality == INT32_C(0x10000)) { ------------------ | Branch (2294:17): [True: 0, False: 0] ------------------ 2295| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 2296| 0| return run_container_create_range(0, INT32_C(0x10000)); 2297| 0| } else { 2298| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 2299| 0| bitset_set_lenrange(bitset->words, min, max - min); 2300| 0| bitset->cardinality = union_cardinality; 2301| 0| return bitset; 2302| 0| } 2303| 0| } 2304| 2.68k| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 2.68k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2304:9): [True: 2.68k, False: 476] ------------------ 2305| 2.68k| array_container_t *array = CAST_array(c); ------------------ | | 54| 2.68k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.68k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2306| | 2307| 2.68k| int32_t nvals_greater = 2308| 2.68k| count_greater(array->array, array->cardinality, (uint16_t)max); 2309| 2.68k| int32_t nvals_less = 2310| 2.68k| count_less(array->array, array->cardinality - nvals_greater, 2311| 2.68k| (uint16_t)min); 2312| 2.68k| int32_t union_cardinality = 2313| 2.68k| nvals_less + (max - min + 1) + nvals_greater; 2314| | 2315| 2.68k| if (union_cardinality == INT32_C(0x10000)) { ------------------ | Branch (2315:17): [True: 664, False: 2.02k] ------------------ 2316| 664| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 664|#define RUN_CONTAINER_TYPE 3 ------------------ 2317| 664| return run_container_create_range(0, INT32_C(0x10000)); 2318| 2.02k| } else if (union_cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (2318:24): [True: 582, False: 1.44k] ------------------ 2319| 582| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 582|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2320| 582| array_container_add_range_nvals(array, min, max, nvals_less, 2321| 582| nvals_greater); 2322| 582| return array; 2323| 1.44k| } else { 2324| 1.44k| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 1.44k|#define BITSET_CONTAINER_TYPE 1 ------------------ 2325| 1.44k| bitset_container_t *bitset = bitset_container_from_array(array); 2326| 1.44k| bitset_set_lenrange(bitset->words, min, max - min); 2327| 1.44k| bitset->cardinality = union_cardinality; 2328| 1.44k| return bitset; 2329| 1.44k| } 2330| 2.68k| } 2331| 476| case RUN_CONTAINER_TYPE: { ------------------ | | 50| 476|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2331:9): [True: 476, False: 2.68k] ------------------ 2332| 476| run_container_t *run = CAST_run(c); ------------------ | | 77| 476|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 476|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2333| | 2334| 476| int32_t nruns_greater = 2335| 476| rle16_count_greater(run->runs, run->n_runs, (uint16_t)max); 2336| 476| int32_t nruns_less = rle16_count_less( 2337| 476| run->runs, run->n_runs - nruns_greater, (uint16_t)min); 2338| | 2339| 476| int32_t run_size_bytes = 2340| 476| (nruns_less + 1 + nruns_greater) * sizeof(rle16_t); 2341| 476| int32_t bitset_size_bytes = 2342| 476| BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t); 2343| | 2344| 476| if (run_size_bytes <= bitset_size_bytes) { ------------------ | Branch (2344:17): [True: 476, False: 0] ------------------ 2345| 476| run_container_add_range_nruns(run, min, max, nruns_less, 2346| 476| nruns_greater); 2347| 476| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 476|#define RUN_CONTAINER_TYPE 3 ------------------ 2348| 476| return run; 2349| 476| } else { 2350| 0| return container_from_run_range(run, min, max, result_type); 2351| 0| } 2352| 476| } 2353| 0| default: ------------------ | Branch (2353:9): [True: 0, False: 3.16k] ------------------ 2354| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2355| 3.16k| } 2356| 3.16k|} roaring.c:container_remove_range: 2371| 20.4k| uint8_t *result_type) { 2372| 20.4k| switch (type) { 2373| 1.16k| case BITSET_CONTAINER_TYPE: { ------------------ | | 48| 1.16k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2373:9): [True: 1.16k, False: 19.2k] ------------------ 2374| 1.16k| bitset_container_t *bitset = CAST_bitset(c); ------------------ | | 52| 1.16k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.16k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2375| | 2376| 1.16k| int32_t result_cardinality = 2377| 1.16k| bitset->cardinality - 2378| 1.16k| bitset_lenrange_cardinality(bitset->words, min, max - min); 2379| | 2380| 1.16k| if (result_cardinality == 0) { ------------------ | Branch (2380:17): [True: 31, False: 1.13k] ------------------ 2381| 31| return NULL; 2382| 1.13k| } else if (result_cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (2382:24): [True: 109, False: 1.02k] ------------------ 2383| 109| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 109|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2384| 109| bitset_reset_range(bitset->words, min, max + 1); 2385| 109| bitset->cardinality = result_cardinality; 2386| 109| return array_container_from_bitset(bitset); 2387| 1.02k| } else { 2388| 1.02k| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 1.02k|#define BITSET_CONTAINER_TYPE 1 ------------------ 2389| 1.02k| bitset_reset_range(bitset->words, min, max + 1); 2390| 1.02k| bitset->cardinality = result_cardinality; 2391| 1.02k| return bitset; 2392| 1.02k| } 2393| 1.16k| } 2394| 3.50k| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 3.50k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2394:9): [True: 3.50k, False: 16.9k] ------------------ 2395| 3.50k| array_container_t *array = CAST_array(c); ------------------ | | 54| 3.50k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 3.50k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2396| | 2397| 3.50k| int32_t nvals_greater = 2398| 3.50k| count_greater(array->array, array->cardinality, (uint16_t)max); 2399| 3.50k| int32_t nvals_less = 2400| 3.50k| count_less(array->array, array->cardinality - nvals_greater, 2401| 3.50k| (uint16_t)min); 2402| 3.50k| int32_t result_cardinality = nvals_less + nvals_greater; 2403| | 2404| 3.50k| if (result_cardinality == 0) { ------------------ | Branch (2404:17): [True: 363, False: 3.14k] ------------------ 2405| 363| return NULL; 2406| 3.14k| } else { 2407| 3.14k| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 3.14k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2408| 3.14k| array_container_remove_range( 2409| 3.14k| array, nvals_less, array->cardinality - result_cardinality); 2410| 3.14k| return array; 2411| 3.14k| } 2412| 3.50k| } 2413| 15.7k| case RUN_CONTAINER_TYPE: { ------------------ | | 50| 15.7k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2413:9): [True: 15.7k, False: 4.67k] ------------------ 2414| 15.7k| run_container_t *run = CAST_run(c); ------------------ | | 77| 15.7k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 15.7k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2415| | 2416| 15.7k| if (run->n_runs == 0) { ------------------ | Branch (2416:17): [True: 0, False: 15.7k] ------------------ 2417| 0| return NULL; 2418| 0| } 2419| 15.7k| if (min <= run_container_minimum(run) && ------------------ | Branch (2419:17): [True: 15.3k, False: 396] ------------------ 2420| 15.3k| max >= run_container_maximum(run)) { ------------------ | Branch (2420:17): [True: 13.9k, False: 1.43k] ------------------ 2421| 13.9k| return NULL; 2422| 13.9k| } 2423| | 2424| 1.83k| run_container_remove_range(run, min, max); 2425| 1.83k| return convert_run_to_efficient_container(run, result_type); 2426| 15.7k| } 2427| 0| default: ------------------ | Branch (2427:9): [True: 0, False: 20.4k] ------------------ 2428| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2429| 20.4k| } 2430| 20.4k|} roaring.c:container_get_cardinality: 244| 368k| uint8_t typecode) { 245| 368k| c = container_unwrap_shared(c, &typecode); 246| 368k| switch (typecode) { ------------------ | Branch (246:13): [True: 368k, False: 0] ------------------ 247| 9.13k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 9.13k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (247:9): [True: 9.13k, False: 359k] ------------------ 248| 9.13k| return bitset_container_cardinality(const_CAST_bitset(c)); ------------------ | | 53| 9.13k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 9.13k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 249| 62.0k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 62.0k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (249:9): [True: 62.0k, False: 306k] ------------------ 250| 62.0k| return array_container_cardinality(const_CAST_array(c)); ------------------ | | 55| 62.0k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 62.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 251| 297k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 297k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (251:9): [True: 297k, False: 71.1k] ------------------ 252| 297k| return run_container_cardinality(const_CAST_run(c)); ------------------ | | 78| 297k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 297k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 253| 368k| } 254| 368k| assert(false); 255| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 256| 0| return 0; // unreached 257| 368k|} roaring.c:get_container_type: 129| 73.8k|static inline uint8_t get_container_type(const container_t *c, uint8_t type) { 130| 73.8k| if (type == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 73.8k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (130:9): [True: 0, False: 73.8k] ------------------ 131| 0| return const_CAST_shared(c)->typecode; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 132| 73.8k| } else { 133| 73.8k| return type; 134| 73.8k| } 135| 73.8k|} roaring.c:get_writable_copy_if_shared: 146| 8.06M| uint8_t *type) { 147| 8.06M| if (*type == SHARED_CONTAINER_TYPE) { // shared, return enclosed container ------------------ | | 51| 8.06M|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (147:9): [True: 0, False: 8.06M] ------------------ 148| 0| return shared_container_extract_copy(CAST_shared(c), type); ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 149| 8.06M| } else { 150| 8.06M| return c; // not shared, so return as-is 151| 8.06M| } 152| 8.06M|} roaring.c:container_add: 492| 7.83M| uint8_t *new_typecode) { 493| 7.83M| c = get_writable_copy_if_shared(c, &typecode); 494| 7.83M| switch (typecode) { 495| 303k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 303k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (495:9): [True: 303k, False: 7.53M] ------------------ 496| 303k| bitset_container_set(CAST_bitset(c), val); ------------------ | | 52| 303k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 303k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 497| 303k| *new_typecode = BITSET_CONTAINER_TYPE; ------------------ | | 48| 303k|#define BITSET_CONTAINER_TYPE 1 ------------------ 498| 303k| return c; 499| 7.12M| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 7.12M|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (499:9): [True: 7.12M, False: 715k] ------------------ 500| 7.12M| array_container_t *ac = CAST_array(c); ------------------ | | 54| 7.12M|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 7.12M|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 501| 7.12M| if (array_container_try_add(ac, val, DEFAULT_MAX_SIZE) != -1) { ------------------ | Branch (501:17): [True: 7.12M, False: 120] ------------------ 502| 7.12M| *new_typecode = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 7.12M|#define ARRAY_CONTAINER_TYPE 2 ------------------ 503| 7.12M| return ac; 504| 7.12M| } else { 505| 120| bitset_container_t *bitset = bitset_container_from_array(ac); 506| 120| bitset_container_add(bitset, val); 507| 120| *new_typecode = BITSET_CONTAINER_TYPE; ------------------ | | 48| 120|#define BITSET_CONTAINER_TYPE 1 ------------------ 508| 120| return bitset; 509| 120| } 510| 7.12M| } break; 511| 412k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 412k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (511:9): [True: 412k, False: 7.42M] ------------------ 512| | // per Java, no container type adjustments are done (revisit?) 513| 412k| run_container_add(CAST_run(c), val); ------------------ | | 77| 412k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 412k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 514| 412k| *new_typecode = RUN_CONTAINER_TYPE; ------------------ | | 50| 412k|#define RUN_CONTAINER_TYPE 3 ------------------ 515| 412k| return c; 516| 0| default: ------------------ | Branch (516:9): [True: 0, False: 7.83M] ------------------ 517| 0| assert(false); 518| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 519| 0| return NULL; 520| 7.83M| } 521| 7.83M|} roaring.c:container_remove: 533| 11.3k| uint8_t *new_typecode) { 534| 11.3k| c = get_writable_copy_if_shared(c, &typecode); 535| 11.3k| switch (typecode) { 536| 2.15k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 2.15k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (536:9): [True: 2.15k, False: 9.19k] ------------------ 537| 2.15k| if (bitset_container_remove(CAST_bitset(c), val)) { ------------------ | | 52| 2.15k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.15k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (537:17): [True: 1.31k, False: 841] ------------------ 538| 1.31k| int card = bitset_container_cardinality(CAST_bitset(c)); ------------------ | | 52| 1.31k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.31k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 539| 1.31k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (539:21): [True: 18, False: 1.30k] ------------------ 540| 18| *new_typecode = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 18|#define ARRAY_CONTAINER_TYPE 2 ------------------ 541| 18| return array_container_from_bitset(CAST_bitset(c)); ------------------ | | 52| 18|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 18|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 542| 18| } 543| 1.31k| } 544| 2.14k| *new_typecode = typecode; 545| 2.14k| return c; 546| 6.04k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 6.04k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (546:9): [True: 6.04k, False: 5.30k] ------------------ 547| 6.04k| *new_typecode = typecode; 548| 6.04k| array_container_remove(CAST_array(c), val); ------------------ | | 54| 6.04k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 6.04k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 549| 6.04k| return c; 550| 3.14k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 3.14k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (550:9): [True: 3.14k, False: 8.20k] ------------------ 551| | // per Java, no container type adjustments are done (revisit?) 552| 3.14k| run_container_remove(CAST_run(c), val); ------------------ | | 77| 3.14k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 3.14k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 553| 3.14k| *new_typecode = RUN_CONTAINER_TYPE; ------------------ | | 50| 3.14k|#define RUN_CONTAINER_TYPE 3 ------------------ 554| 3.14k| return c; 555| 0| default: ------------------ | Branch (555:9): [True: 0, False: 11.3k] ------------------ 556| 0| assert(false); 557| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 558| 0| return NULL; 559| 11.3k| } 560| 11.3k|} roaring.c:container_nonzero_cardinality: 437| 168k| uint8_t typecode) { 438| 168k| c = container_unwrap_shared(c, &typecode); 439| 168k| switch (typecode) { ------------------ | Branch (439:13): [True: 168k, False: 0] ------------------ 440| 6.54k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 6.54k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (440:9): [True: 6.54k, False: 161k] ------------------ 441| 6.54k| return bitset_container_const_nonzero_cardinality( 442| 6.54k| const_CAST_bitset(c)); ------------------ | | 53| 6.54k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 6.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 443| 100k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 100k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (443:9): [True: 100k, False: 68.1k] ------------------ 444| 100k| return array_container_nonzero_cardinality(const_CAST_array(c)); ------------------ | | 55| 100k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 100k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 445| 61.6k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 61.6k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (445:9): [True: 61.6k, False: 106k] ------------------ 446| 61.6k| return run_container_nonzero_cardinality(const_CAST_run(c)); ------------------ | | 78| 61.6k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 61.6k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 447| 168k| } 448| 168k| assert(false); 449| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 450| 0| return 0; // unreached 451| 168k|} roaring.c:container_and: 728| 6.12k| uint8_t *result_type) { 729| 6.12k| c1 = container_unwrap_shared(c1, &type1); 730| 6.12k| c2 = container_unwrap_shared(c2, &type2); 731| 6.12k| container_t *result = NULL; 732| 6.12k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.12k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 733| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (733:9): [True: 0, False: 6.12k] ------------------ 734| 0| *result_type = 735| 0| bitset_bitset_container_intersection( ------------------ | Branch (735:17): [True: 0, False: 0] ------------------ 736| 0| const_CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 737| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 738| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 739| 0| return result; 740| | 741| 2.30k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.30k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (741:9): [True: 2.30k, False: 3.82k] ------------------ 742| 2.30k| result = array_container_create(); 743| 2.30k| array_container_intersection( 744| 2.30k| const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 54| 2.30k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 745| 2.30k| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 746| 2.30k| return result; 747| | 748| 690| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 690| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (748:9): [True: 690, False: 5.43k] ------------------ 749| 690| result = run_container_create(); 750| 690| run_container_intersection(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_intersection(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 751| 690| CAST_run(result)); ------------------ | | 77| 690|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 752| 690| return convert_run_to_efficient_container_and_free(CAST_run(result), ------------------ | | 77| 690|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 753| 690| result_type); 754| | 755| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (755:9): [True: 0, False: 6.12k] ------------------ 756| 0| result = array_container_create(); 757| 0| array_bitset_container_intersection(const_CAST_array(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 758| 0| const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 759| 0| CAST_array(result)); ------------------ | | 54| 0|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 760| 0| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 761| 0| return result; 762| | 763| 531| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 531| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 531|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (763:9): [True: 531, False: 5.59k] ------------------ 764| 531| result = array_container_create(); 765| 531| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 ------------------ 766| 531| array_bitset_container_intersection(const_CAST_array(c1), ------------------ | | 55| 531|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 767| 531| const_CAST_bitset(c2), ------------------ | | 53| 531|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 768| 531| CAST_array(result)); ------------------ | | 54| 531|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 769| 531| return result; 770| | 771| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (771:9): [True: 0, False: 6.12k] ------------------ 772| 0| *result_type = 773| 0| run_bitset_container_intersection( ------------------ | Branch (773:17): [True: 0, False: 0] ------------------ 774| 0| const_CAST_run(c2), const_CAST_bitset(c1), &result) ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c2), const_CAST_bitset(c1), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 775| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 776| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 777| 0| return result; 778| | 779| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (779:9): [True: 613, False: 5.51k] ------------------ 780| 613| *result_type = 781| 613| run_bitset_container_intersection( ------------------ | Branch (781:17): [True: 0, False: 613] ------------------ 782| 613| const_CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 783| 613| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 784| 613| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.22k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 785| 613| return result; 786| | 787| 783| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 783| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 783|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (787:9): [True: 783, False: 5.34k] ------------------ 788| 783| result = array_container_create(); 789| 783| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 ------------------ 790| 783| array_run_container_intersection( 791| 783| const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 55| 783|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 78| 783|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 54| 783|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 792| 783| return result; 793| | 794| 1.20k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 1.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.20k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (794:9): [True: 1.20k, False: 4.92k] ------------------ 795| 1.20k| result = array_container_create(); 796| 1.20k| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 797| 1.20k| array_run_container_intersection( 798| 1.20k| const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 55| 1.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 78| 1.20k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 54| 1.20k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 799| 1.20k| return result; 800| | 801| 0| default: ------------------ | Branch (801:9): [True: 0, False: 6.12k] ------------------ 802| 0| assert(false); 803| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 804| 0| return NULL; 805| 6.12k| } 806| 6.12k|} roaring.c:container_iand: 923| 63.2k| uint8_t *result_type) { 924| 63.2k| c1 = get_writable_copy_if_shared(c1, &type1); 925| 63.2k| c2 = container_unwrap_shared(c2, &type2); 926| 63.2k| container_t *result = NULL; 927| 63.2k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 63.2k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 928| 1.40k| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 1.40k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 1.40k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 1.40k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (928:9): [True: 1.40k, False: 61.8k] ------------------ 929| 1.40k| *result_type = bitset_bitset_container_intersection_inplace( ------------------ | Branch (929:28): [True: 1.40k, False: 0] ------------------ 930| 1.40k| CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 52| 1.40k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.40k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 1.40k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.40k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 931| 1.40k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 1.40k|#define BITSET_CONTAINER_TYPE 1 ------------------ 932| 1.40k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.40k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 933| 1.40k| return result; 934| | 935| 4.57k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 4.57k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 4.57k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 4.57k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (935:9): [True: 4.57k, False: 58.7k] ------------------ 936| 4.57k| array_container_intersection_inplace(CAST_array(c1), ------------------ | | 54| 4.57k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 4.57k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 937| 4.57k| const_CAST_array(c2)); ------------------ | | 55| 4.57k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 4.57k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 938| 4.57k| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 4.57k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 939| 4.57k| return c1; 940| | 941| 56.0k| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 56.0k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 56.0k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 56.0k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (941:9): [True: 56.0k, False: 7.22k] ------------------ 942| 56.0k| result = run_container_create(); 943| 56.0k| run_container_intersection(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 56.0k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 56.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_intersection(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 56.0k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 56.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 944| 56.0k| CAST_run(result)); ------------------ | | 77| 56.0k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 56.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 945| | // as of January 2016, Java code used non-in-place intersection for 946| | // two runcontainers 947| 56.0k| return convert_run_to_efficient_container_and_free(CAST_run(result), ------------------ | | 77| 56.0k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 56.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 948| 56.0k| result_type); 949| | 950| 54| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 54| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 54|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 54|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (950:9): [True: 54, False: 63.2k] ------------------ 951| | // c1 is a bitmap so no inplace possible 952| 54| result = array_container_create(); 953| 54| array_bitset_container_intersection(const_CAST_array(c2), ------------------ | | 55| 54|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 54|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 954| 54| const_CAST_bitset(c1), ------------------ | | 53| 54|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 54|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 955| 54| CAST_array(result)); ------------------ | | 54| 54|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 54|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 956| 54| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 54|#define ARRAY_CONTAINER_TYPE 2 ------------------ 957| 54| return result; 958| | 959| 0| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (959:9): [True: 0, False: 63.2k] ------------------ 960| 0| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 961| 0| array_bitset_container_intersection( 962| 0| const_CAST_array(c1), const_CAST_bitset(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_bitset(c2), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 963| 0| CAST_array(c1)); // result is allowed to be same as c1 ------------------ | | 54| 0|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 964| 0| return c1; 965| | 966| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (966:9): [True: 0, False: 63.2k] ------------------ 967| | // will attempt in-place computation 968| 0| *result_type = run_bitset_container_intersection( ------------------ | Branch (968:28): [True: 0, False: 0] ------------------ 969| 0| const_CAST_run(c2), const_CAST_bitset(c1), &c1) ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c2), const_CAST_bitset(c1), &c1) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 970| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 971| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 972| 0| return c1; 973| | 974| 0| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (974:9): [True: 0, False: 63.2k] ------------------ 975| 0| *result_type = 976| 0| run_bitset_container_intersection( ------------------ | Branch (976:17): [True: 0, False: 0] ------------------ 977| 0| const_CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 978| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 979| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 980| 0| return result; 981| | 982| 198| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 198| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 198|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 198|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (982:9): [True: 198, False: 63.0k] ------------------ 983| 198| result = array_container_create(); 984| 198| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 198|#define ARRAY_CONTAINER_TYPE 2 ------------------ 985| 198| array_run_container_intersection( 986| 198| const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 55| 198|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 198|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 78| 198|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 198|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2), CAST_array(result)); ------------------ | | 54| 198|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 198|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 987| 198| return result; 988| | 989| 999| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 999| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 999|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 999|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (989:9): [True: 999, False: 62.2k] ------------------ 990| 999| result = array_container_create(); 991| 999| *result_type = ARRAY_CONTAINER_TYPE; // never bitset ------------------ | | 49| 999|#define ARRAY_CONTAINER_TYPE 2 ------------------ 992| 999| array_run_container_intersection( 993| 999| const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 55| 999|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 999|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 78| 999|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 999|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1), CAST_array(result)); ------------------ | | 54| 999|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 999|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 994| 999| return result; 995| | 996| 0| default: ------------------ | Branch (996:9): [True: 0, False: 63.2k] ------------------ 997| 0| assert(false); 998| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 999| 0| return NULL; 1000| 63.2k| } 1001| 63.2k|} roaring.c:container_or: 1010| 6.12k| uint8_t *result_type) { 1011| 6.12k| c1 = container_unwrap_shared(c1, &type1); 1012| 6.12k| c2 = container_unwrap_shared(c2, &type2); 1013| 6.12k| container_t *result = NULL; 1014| 6.12k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.12k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1015| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1015:9): [True: 0, False: 6.12k] ------------------ 1016| 0| result = bitset_container_create(); 1017| 0| bitset_container_or(const_CAST_bitset(c1), const_CAST_bitset(c2), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ bitset_container_or(const_CAST_bitset(c1), const_CAST_bitset(c2), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1018| 0| CAST_bitset(result)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1019| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1020| 0| return result; 1021| | 1022| 2.30k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.30k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1022:9): [True: 2.30k, False: 3.82k] ------------------ 1023| 2.30k| *result_type = 1024| 2.30k| array_array_container_union(const_CAST_array(c1), ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1024:17): [True: 54, False: 2.25k] ------------------ 1025| 2.30k| const_CAST_array(c2), &result) ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1026| 2.30k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 54|#define BITSET_CONTAINER_TYPE 1 ------------------ 1027| 2.30k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 4.55k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1028| 2.30k| return result; 1029| | 1030| 690| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 690| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1030:9): [True: 690, False: 5.43k] ------------------ 1031| 690| result = run_container_create(); 1032| 690| run_container_union(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_union(const_CAST_run(c1), const_CAST_run(c2), ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1033| 690| CAST_run(result)); ------------------ | | 77| 690|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1034| 690| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 690|#define RUN_CONTAINER_TYPE 3 ------------------ 1035| | // todo: could be optimized since will never convert to array 1036| 690| result = convert_run_to_efficient_container_and_free( 1037| 690| CAST_run(result), result_type); ------------------ | | 77| 690|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1038| 690| return result; 1039| | 1040| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1040:9): [True: 0, False: 6.12k] ------------------ 1041| 0| result = bitset_container_create(); 1042| 0| array_bitset_container_union(const_CAST_array(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1043| 0| const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1044| 0| CAST_bitset(result)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1045| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1046| 0| return result; 1047| | 1048| 531| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 531| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 531|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1048:9): [True: 531, False: 5.59k] ------------------ 1049| 531| result = bitset_container_create(); 1050| 531| array_bitset_container_union(const_CAST_array(c1), ------------------ | | 55| 531|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1051| 531| const_CAST_bitset(c2), ------------------ | | 53| 531|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1052| 531| CAST_bitset(result)); ------------------ | | 52| 531|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1053| 531| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 531|#define BITSET_CONTAINER_TYPE 1 ------------------ 1054| 531| return result; 1055| | 1056| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1056:9): [True: 0, False: 6.12k] ------------------ 1057| 0| if (run_container_is_full(const_CAST_run(c2))) { ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1057:17): [True: 0, False: 0] ------------------ 1058| 0| result = run_container_create(); 1059| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 1060| 0| run_container_copy(const_CAST_run(c2), CAST_run(result)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_copy(const_CAST_run(c2), CAST_run(result)); ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1061| 0| return result; 1062| 0| } 1063| 0| result = bitset_container_create(); 1064| 0| run_bitset_container_union( 1065| 0| const_CAST_run(c2), const_CAST_bitset(c1), CAST_bitset(result)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c2), const_CAST_bitset(c1), CAST_bitset(result)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c2), const_CAST_bitset(c1), CAST_bitset(result)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1066| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1067| 0| return result; 1068| | 1069| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1069:9): [True: 613, False: 5.51k] ------------------ 1070| 613| if (run_container_is_full(const_CAST_run(c1))) { ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1070:17): [True: 0, False: 613] ------------------ 1071| 0| result = run_container_create(); 1072| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 1073| 0| run_container_copy(const_CAST_run(c1), CAST_run(result)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_copy(const_CAST_run(c1), CAST_run(result)); ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1074| 0| return result; 1075| 0| } 1076| 613| result = bitset_container_create(); 1077| 613| run_bitset_container_union( 1078| 613| const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 52| 613|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1079| 613| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 613|#define BITSET_CONTAINER_TYPE 1 ------------------ 1080| 613| return result; 1081| | 1082| 783| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 783| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 783|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1082:9): [True: 783, False: 5.34k] ------------------ 1083| 783| result = run_container_create(); 1084| 783| array_run_container_union(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 55| 783|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_run_container_union(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 78| 783|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1085| 783| CAST_run(result)); ------------------ | | 77| 783|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1086| 783| result = convert_run_to_efficient_container_and_free( 1087| 783| CAST_run(result), result_type); ------------------ | | 77| 783|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1088| 783| return result; 1089| | 1090| 1.20k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 1.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.20k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1090:9): [True: 1.20k, False: 4.92k] ------------------ 1091| 1.20k| result = run_container_create(); 1092| 1.20k| array_run_container_union(const_CAST_array(c2), const_CAST_run(c1), ------------------ | | 55| 1.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_run_container_union(const_CAST_array(c2), const_CAST_run(c1), ------------------ | | 78| 1.20k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1093| 1.20k| CAST_run(result)); ------------------ | | 77| 1.20k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1094| 1.20k| result = convert_run_to_efficient_container_and_free( 1095| 1.20k| CAST_run(result), result_type); ------------------ | | 77| 1.20k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1096| 1.20k| return result; 1097| | 1098| 0| default: ------------------ | Branch (1098:9): [True: 0, False: 6.12k] ------------------ 1099| 0| assert(false); 1100| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1101| 0| return NULL; // unreached 1102| 6.12k| } 1103| 6.12k|} roaring.c:container_is_full: 262| 7.00k|static inline bool container_is_full(const container_t *c, uint8_t typecode) { 263| 7.00k| c = container_unwrap_shared(c, &typecode); 264| 7.00k| switch (typecode) { ------------------ | Branch (264:13): [True: 7.00k, False: 0] ------------------ 265| 0| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (265:9): [True: 0, False: 7.00k] ------------------ 266| 0| return bitset_container_cardinality(const_CAST_bitset(c)) == ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 267| 0| (1 << 16); 268| 3.70k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.70k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (268:9): [True: 3.70k, False: 3.30k] ------------------ 269| 3.70k| return array_container_cardinality(const_CAST_array(c)) == ------------------ | | 55| 3.70k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.70k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 270| 3.70k| (1 << 16); 271| 3.30k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 3.30k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (271:9): [True: 3.30k, False: 3.70k] ------------------ 272| 3.30k| return run_container_is_full(const_CAST_run(c)); ------------------ | | 78| 3.30k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 3.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 273| 7.00k| } 274| 7.00k| assert(false); 275| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 276| 0| return 0; // unreached 277| 7.00k|} roaring.c:container_ior: 1228| 6.24k| uint8_t *result_type) { 1229| 6.24k| c1 = get_writable_copy_if_shared(c1, &type1); 1230| 6.24k| c2 = container_unwrap_shared(c2, &type2); 1231| 6.24k| container_t *result = NULL; 1232| 6.24k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.24k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1233| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1233:9): [True: 0, False: 6.24k] ------------------ 1234| 0| bitset_container_or(const_CAST_bitset(c1), const_CAST_bitset(c2), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ bitset_container_or(const_CAST_bitset(c1), const_CAST_bitset(c2), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1235| 0| CAST_bitset(c1)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1236| 0|#ifdef OR_BITSET_CONVERSION_TO_FULL 1237| 0| if (CAST_bitset(c1)->cardinality == (1 << 16)) { // we convert ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1237:17): [True: 0, False: 0] ------------------ 1238| 0| result = run_container_create_range(0, (1 << 16)); 1239| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 1240| 0| return result; 1241| 0| } 1242| 0|#endif 1243| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1244| 0| return c1; 1245| | 1246| 2.47k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.47k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.47k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.47k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1246:9): [True: 2.47k, False: 3.76k] ------------------ 1247| 2.47k| *result_type = array_array_container_inplace_union( ------------------ | Branch (1247:28): [True: 0, False: 2.47k] ------------------ 1248| 2.47k| CAST_array(c1), const_CAST_array(c2), &result) ------------------ | | 54| 2.47k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.47k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_array(c1), const_CAST_array(c2), &result) ------------------ | | 55| 2.47k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.47k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1249| 2.47k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1250| 2.47k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 4.95k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1251| 2.47k| if ((result == NULL) && (*result_type == ARRAY_CONTAINER_TYPE)) { ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (1251:17): [True: 0, False: 2.47k] | Branch (1251:37): [True: 0, False: 0] ------------------ 1252| 0| return c1; // the computation was done in-place! 1253| 0| } 1254| 2.47k| return result; 1255| | 1256| 1.67k| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 1.67k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.67k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.67k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1256:9): [True: 1.67k, False: 4.57k] ------------------ 1257| 1.67k| run_container_union_inplace(CAST_run(c1), const_CAST_run(c2)); ------------------ | | 77| 1.67k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.67k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_union_inplace(CAST_run(c1), const_CAST_run(c2)); ------------------ | | 78| 1.67k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.67k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1258| 1.67k| return convert_run_to_efficient_container(CAST_run(c1), ------------------ | | 77| 1.67k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.67k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1259| 1.67k| result_type); 1260| | 1261| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1261:9): [True: 0, False: 6.24k] ------------------ 1262| 0| array_bitset_container_union( 1263| 0| const_CAST_array(c2), const_CAST_bitset(c1), CAST_bitset(c1)); ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_bitset(c1), CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_bitset(c1), CAST_bitset(c1)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1264| 0| *result_type = BITSET_CONTAINER_TYPE; // never array ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1265| 0| return c1; 1266| | 1267| 585| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 585| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 585|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 585|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1267:9): [True: 585, False: 5.66k] ------------------ 1268| | // c1 is an array, so no in-place possible 1269| 585| result = bitset_container_create(); 1270| 585| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 585|#define BITSET_CONTAINER_TYPE 1 ------------------ 1271| 585| array_bitset_container_union(const_CAST_array(c1), ------------------ | | 55| 585|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 585|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1272| 585| const_CAST_bitset(c2), ------------------ | | 53| 585|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 585|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1273| 585| CAST_bitset(result)); ------------------ | | 52| 585|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 585|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1274| 585| return result; 1275| | 1276| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1276:9): [True: 0, False: 6.24k] ------------------ 1277| 0| if (run_container_is_full(const_CAST_run(c2))) { ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1277:17): [True: 0, False: 0] ------------------ 1278| 0| result = run_container_create(); 1279| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 1280| 0| run_container_copy(const_CAST_run(c2), CAST_run(result)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ run_container_copy(const_CAST_run(c2), CAST_run(result)); ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1281| 0| return result; 1282| 0| } 1283| 0| run_bitset_container_union(const_CAST_run(c2), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1284| 0| const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1285| 0| CAST_bitset(c1)); // allowed ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1286| 0| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1287| 0| return c1; 1288| | 1289| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1289:9): [True: 613, False: 5.63k] ------------------ 1290| 613| if (run_container_is_full(const_CAST_run(c1))) { ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1290:17): [True: 0, False: 613] ------------------ 1291| 0| *result_type = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 1292| 0| return c1; 1293| 0| } 1294| 613| result = bitset_container_create(); 1295| 613| run_bitset_container_union( 1296| 613| const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2), CAST_bitset(result)); ------------------ | | 52| 613|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1297| 613| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 613|#define BITSET_CONTAINER_TYPE 1 ------------------ 1298| 613| return result; 1299| | 1300| 636| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 636| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 636|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 636|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1300:9): [True: 636, False: 5.61k] ------------------ 1301| 636| result = run_container_create(); 1302| 636| array_run_container_union(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 55| 636|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 636|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_run_container_union(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 78| 636|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 636|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1303| 636| CAST_run(result)); ------------------ | | 77| 636|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 636|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1304| 636| result = convert_run_to_efficient_container_and_free( 1305| 636| CAST_run(result), result_type); ------------------ | | 77| 636|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 636|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1306| 636| return result; 1307| | 1308| 259| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 259| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 259|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 259|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1308:9): [True: 259, False: 5.98k] ------------------ 1309| 259| array_run_container_inplace_union(const_CAST_array(c2), ------------------ | | 55| 259|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 259|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1310| 259| CAST_run(c1)); ------------------ | | 77| 259|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 259|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1311| 259| c1 = convert_run_to_efficient_container(CAST_run(c1), result_type); ------------------ | | 77| 259|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 259|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1312| 259| return c1; 1313| | 1314| 0| default: ------------------ | Branch (1314:9): [True: 0, False: 6.24k] ------------------ 1315| 0| assert(false); 1316| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1317| 0| return NULL; 1318| 6.24k| } 1319| 6.24k|} roaring.c:container_xor: 1451| 6.12k| uint8_t *result_type) { 1452| 6.12k| c1 = container_unwrap_shared(c1, &type1); 1453| 6.12k| c2 = container_unwrap_shared(c2, &type2); 1454| 6.12k| container_t *result = NULL; 1455| 6.12k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.12k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1456| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1456:9): [True: 0, False: 6.12k] ------------------ 1457| 0| *result_type = 1458| 0| bitset_bitset_container_xor(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1458:17): [True: 0, False: 0] ------------------ 1459| 0| const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1460| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1461| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1462| 0| return result; 1463| | 1464| 2.30k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.30k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1464:9): [True: 2.30k, False: 3.82k] ------------------ 1465| 2.30k| *result_type = 1466| 2.30k| array_array_container_xor(const_CAST_array(c1), ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1466:17): [True: 12, False: 2.29k] ------------------ 1467| 2.30k| const_CAST_array(c2), &result) ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1468| 2.30k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 12|#define BITSET_CONTAINER_TYPE 1 ------------------ 1469| 2.30k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 4.59k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1470| 2.30k| return result; 1471| | 1472| 690| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 690| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1472:9): [True: 690, False: 5.43k] ------------------ 1473| 690| *result_type = (uint8_t)run_run_container_xor( 1474| 690| const_CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1475| 690| return result; 1476| | 1477| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1477:9): [True: 0, False: 6.12k] ------------------ 1478| 0| *result_type = 1479| 0| array_bitset_container_xor(const_CAST_array(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1479:17): [True: 0, False: 0] ------------------ 1480| 0| const_CAST_bitset(c1), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1481| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1482| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1483| 0| return result; 1484| | 1485| 531| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 531| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 531|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1485:9): [True: 531, False: 5.59k] ------------------ 1486| 531| *result_type = 1487| 531| array_bitset_container_xor(const_CAST_array(c1), ------------------ | | 55| 531|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1487:17): [True: 483, False: 48] ------------------ 1488| 531| const_CAST_bitset(c2), &result) ------------------ | | 53| 531|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1489| 531| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 483|#define BITSET_CONTAINER_TYPE 1 ------------------ 1490| 531| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 579|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1491| 531| return result; 1492| | 1493| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1493:9): [True: 0, False: 6.12k] ------------------ 1494| 0| *result_type = 1495| 0| run_bitset_container_xor(const_CAST_run(c2), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1495:17): [True: 0, False: 0] ------------------ 1496| 0| const_CAST_bitset(c1), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1497| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1498| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1499| 0| return result; 1500| | 1501| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1501:9): [True: 613, False: 5.51k] ------------------ 1502| 613| *result_type = 1503| 613| run_bitset_container_xor(const_CAST_run(c1), ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1503:17): [True: 564, False: 49] ------------------ 1504| 613| const_CAST_bitset(c2), &result) ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1505| 613| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 564|#define BITSET_CONTAINER_TYPE 1 ------------------ 1506| 613| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 662|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1507| 613| return result; 1508| | 1509| 783| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 783| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 783|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1509:9): [True: 783, False: 5.34k] ------------------ 1510| 783| *result_type = (uint8_t)array_run_container_xor( 1511| 783| const_CAST_array(c1), const_CAST_run(c2), &result); ------------------ | | 55| 783|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2), &result); ------------------ | | 78| 783|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1512| 783| return result; 1513| | 1514| 1.20k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 1.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.20k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1514:9): [True: 1.20k, False: 4.92k] ------------------ 1515| 1.20k| *result_type = (uint8_t)array_run_container_xor( 1516| 1.20k| const_CAST_array(c2), const_CAST_run(c1), &result); ------------------ | | 55| 1.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1), &result); ------------------ | | 78| 1.20k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1517| 1.20k| return result; 1518| | 1519| 0| default: ------------------ | Branch (1519:9): [True: 0, False: 6.12k] ------------------ 1520| 0| assert(false); 1521| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1522| 0| return NULL; // unreached 1523| 6.12k| } 1524| 6.12k|} roaring.c:container_ixor: 1667| 61.7k| uint8_t *result_type) { 1668| 61.7k| c1 = get_writable_copy_if_shared(c1, &type1); 1669| 61.7k| c2 = container_unwrap_shared(c2, &type2); 1670| 61.7k| container_t *result = NULL; 1671| 61.7k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 61.7k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1672| 1.45k| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 1.45k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 1.45k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 1.45k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1672:9): [True: 1.45k, False: 60.3k] ------------------ 1673| 1.45k| *result_type = bitset_bitset_container_ixor( ------------------ | Branch (1673:28): [True: 0, False: 1.45k] ------------------ 1674| 1.45k| CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 52| 1.45k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.45k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 1.45k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.45k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1675| 1.45k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1676| 1.45k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 2.91k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1677| 1.45k| return result; 1678| | 1679| 4.78k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 4.78k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 4.78k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 4.78k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1679:9): [True: 4.78k, False: 57.0k] ------------------ 1680| 4.78k| *result_type = array_array_container_ixor( ------------------ | Branch (1680:28): [True: 0, False: 4.78k] ------------------ 1681| 4.78k| CAST_array(c1), const_CAST_array(c2), &result) ------------------ | | 54| 4.78k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 4.78k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_array(c1), const_CAST_array(c2), &result) ------------------ | | 55| 4.78k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 4.78k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1682| 4.78k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1683| 4.78k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 9.56k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1684| 4.78k| return result; 1685| | 1686| 55.0k| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 55.0k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 55.0k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 55.0k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1686:9): [True: 55.0k, False: 6.73k] ------------------ 1687| 55.0k| *result_type = (uint8_t)run_run_container_ixor( 1688| 55.0k| CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 77| 55.0k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 55.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 55.0k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 55.0k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1689| 55.0k| return result; 1690| | 1691| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1691:9): [True: 0, False: 61.7k] ------------------ 1692| 0| *result_type = bitset_array_container_ixor( ------------------ | Branch (1692:28): [True: 0, False: 0] ------------------ 1693| 0| CAST_bitset(c1), const_CAST_array(c2), &result) ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_array(c2), &result) ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1694| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1695| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1696| 0| return result; 1697| | 1698| 88| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 88| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 88|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 88|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1698:9): [True: 88, False: 61.7k] ------------------ 1699| 88| *result_type = array_bitset_container_ixor( ------------------ | Branch (1699:28): [True: 0, False: 88] ------------------ 1700| 88| CAST_array(c1), const_CAST_bitset(c2), &result) ------------------ | | 54| 88|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 88|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_array(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 88|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 88|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1701| 88| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1702| 88| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 176|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1703| 88| return result; 1704| | 1705| 63| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 63| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 63|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 63|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1705:9): [True: 63, False: 61.7k] ------------------ 1706| 63| *result_type = bitset_run_container_ixor( ------------------ | Branch (1706:28): [True: 0, False: 63] ------------------ 1707| 63| CAST_bitset(c1), const_CAST_run(c2), &result) ------------------ | | 52| 63|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 63|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_run(c2), &result) ------------------ | | 78| 63|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 63|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1708| 63| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1709| 63| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 126|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1710| | 1711| 63| return result; 1712| | 1713| 22| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 22| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 22|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 22|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1713:9): [True: 22, False: 61.7k] ------------------ 1714| 22| *result_type = run_bitset_container_ixor( ------------------ | Branch (1714:28): [True: 0, False: 22] ------------------ 1715| 22| CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 77| 22|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 22|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 22|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 22|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1716| 22| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1717| 22| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 44|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1718| 22| return result; 1719| | 1720| 262| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 262| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 262|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 262|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1720:9): [True: 262, False: 61.5k] ------------------ 1721| 262| *result_type = (uint8_t)array_run_container_ixor( 1722| 262| CAST_array(c1), const_CAST_run(c2), &result); ------------------ | | 54| 262|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 262|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_array(c1), const_CAST_run(c2), &result); ------------------ | | 78| 262|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 262|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1723| 262| return result; 1724| | 1725| 58| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 58| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 58|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 58|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1725:9): [True: 58, False: 61.7k] ------------------ 1726| 58| *result_type = (uint8_t)run_array_container_ixor( 1727| 58| CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 77| 58|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 58|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 55| 58|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 58|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1728| 58| return result; 1729| | 1730| 0| default: ------------------ | Branch (1730:9): [True: 0, False: 61.7k] ------------------ 1731| 0| assert(false); 1732| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1733| 0| return NULL; 1734| 61.7k| } 1735| 61.7k|} roaring.c:container_andnot: 1787| 6.12k| uint8_t *result_type) { 1788| 6.12k| c1 = container_unwrap_shared(c1, &type1); 1789| 6.12k| c2 = container_unwrap_shared(c2, &type2); 1790| 6.12k| container_t *result = NULL; 1791| 6.12k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.12k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1792| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1792:9): [True: 0, False: 6.12k] ------------------ 1793| 0| *result_type = 1794| 0| bitset_bitset_container_andnot(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1794:17): [True: 0, False: 0] ------------------ 1795| 0| const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1796| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1797| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1798| 0| return result; 1799| | 1800| 2.30k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.30k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1800:9): [True: 2.30k, False: 3.82k] ------------------ 1801| 2.30k| result = array_container_create(); 1802| 2.30k| array_array_container_andnot( 1803| 2.30k| const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_array(c2), CAST_array(result)); ------------------ | | 54| 2.30k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1804| 2.30k| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1805| 2.30k| return result; 1806| | 1807| 690| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 690| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1807:9): [True: 690, False: 5.43k] ------------------ 1808| 690| if (run_container_is_full(const_CAST_run(c2))) { ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1808:17): [True: 14, False: 676] ------------------ 1809| 14| result = array_container_create(); 1810| 14| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 14|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1811| 14| return result; 1812| 14| } 1813| 676| *result_type = (uint8_t)run_run_container_andnot( 1814| 676| const_CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 676|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 676|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 676|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 676|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1815| 676| return result; 1816| | 1817| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1817:9): [True: 0, False: 6.12k] ------------------ 1818| 0| *result_type = 1819| 0| bitset_array_container_andnot(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1819:17): [True: 0, False: 0] ------------------ 1820| 0| const_CAST_array(c2), &result) ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1821| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1822| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1823| 0| return result; 1824| | 1825| 531| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 531| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 531|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1825:9): [True: 531, False: 5.59k] ------------------ 1826| 531| result = array_container_create(); 1827| 531| array_bitset_container_andnot(const_CAST_array(c1), ------------------ | | 55| 531|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1828| 531| const_CAST_bitset(c2), ------------------ | | 53| 531|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1829| 531| CAST_array(result)); ------------------ | | 54| 531|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1830| 531| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1831| 531| return result; 1832| | 1833| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1833:9): [True: 0, False: 6.12k] ------------------ 1834| 0| if (run_container_is_full(const_CAST_run(c2))) { ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1834:17): [True: 0, False: 0] ------------------ 1835| 0| result = array_container_create(); 1836| 0| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1837| 0| return result; 1838| 0| } 1839| 0| *result_type = 1840| 0| bitset_run_container_andnot(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1840:17): [True: 0, False: 0] ------------------ 1841| 0| const_CAST_run(c2), &result) ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1842| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1843| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1844| 0| return result; 1845| | 1846| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1846:9): [True: 613, False: 5.51k] ------------------ 1847| 613| *result_type = 1848| 613| run_bitset_container_andnot(const_CAST_run(c1), ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1848:17): [True: 0, False: 613] ------------------ 1849| 613| const_CAST_bitset(c2), &result) ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1850| 613| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1851| 613| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.22k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1852| 613| return result; 1853| | 1854| 783| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 783| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 783|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1854:9): [True: 783, False: 5.34k] ------------------ 1855| 783| if (run_container_is_full(const_CAST_run(c2))) { ------------------ | | 78| 783|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (1855:17): [True: 24, False: 759] ------------------ 1856| 24| result = array_container_create(); 1857| 24| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 24|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1858| 24| return result; 1859| 24| } 1860| 759| result = array_container_create(); 1861| 759| array_run_container_andnot(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 55| 759|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 759|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_run_container_andnot(const_CAST_array(c1), const_CAST_run(c2), ------------------ | | 78| 759|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 759|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1862| 759| CAST_array(result)); ------------------ | | 54| 759|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 759|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1863| 759| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 759|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1864| 759| return result; 1865| | 1866| 1.20k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 1.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.20k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1866:9): [True: 1.20k, False: 4.92k] ------------------ 1867| 1.20k| *result_type = (uint8_t)run_array_container_andnot( 1868| 1.20k| const_CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 78| 1.20k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 55| 1.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1869| 1.20k| return result; 1870| | 1871| 0| default: ------------------ | Branch (1871:9): [True: 0, False: 6.12k] ------------------ 1872| 0| assert(false); 1873| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1874| 0| return NULL; // unreached 1875| 6.12k| } 1876| 6.12k|} roaring.c:container_iandnot: 1891| 5.81k| uint8_t *result_type) { 1892| 5.81k| c1 = get_writable_copy_if_shared(c1, &type1); 1893| 5.81k| c2 = container_unwrap_shared(c2, &type2); 1894| 5.81k| container_t *result = NULL; 1895| 5.81k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 5.81k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 1896| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1896:9): [True: 0, False: 5.81k] ------------------ 1897| 0| *result_type = bitset_bitset_container_iandnot( ------------------ | Branch (1897:28): [True: 0, False: 0] ------------------ 1898| 0| CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1899| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1900| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1901| 0| return result; 1902| | 1903| 3.44k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 3.44k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 3.44k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 3.44k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1903:9): [True: 3.44k, False: 2.36k] ------------------ 1904| 3.44k| array_array_container_iandnot(CAST_array(c1), const_CAST_array(c2)); ------------------ | | 54| 3.44k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 3.44k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_array_container_iandnot(CAST_array(c1), const_CAST_array(c2)); ------------------ | | 55| 3.44k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.44k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1905| 3.44k| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 3.44k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1906| 3.44k| return c1; 1907| | 1908| 463| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 463| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 463|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 463|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1908:9): [True: 463, False: 5.34k] ------------------ 1909| 463| *result_type = (uint8_t)run_run_container_iandnot( 1910| 463| CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 77| 463|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 463|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_run(c2), &result); ------------------ | | 78| 463|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 463|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1911| 463| return result; 1912| | 1913| 1.11k| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 1.11k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 1.11k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.11k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1913:9): [True: 1.11k, False: 4.69k] ------------------ 1914| 1.11k| *result_type = bitset_array_container_iandnot( ------------------ | Branch (1914:28): [True: 947, False: 167] ------------------ 1915| 1.11k| CAST_bitset(c1), const_CAST_array(c2), &result) ------------------ | | 52| 1.11k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.11k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_array(c2), &result) ------------------ | | 55| 1.11k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.11k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1916| 1.11k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 947|#define BITSET_CONTAINER_TYPE 1 ------------------ 1917| 1.11k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.28k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1918| 1.11k| return result; 1919| | 1920| 0| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1920:9): [True: 0, False: 5.81k] ------------------ 1921| 0| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1922| 0| array_bitset_container_iandnot(CAST_array(c1), ------------------ | | 54| 0|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1923| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1924| 0| return c1; 1925| | 1926| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1926:9): [True: 0, False: 5.81k] ------------------ 1927| 0| *result_type = bitset_run_container_iandnot( ------------------ | Branch (1927:28): [True: 0, False: 0] ------------------ 1928| 0| CAST_bitset(c1), const_CAST_run(c2), &result) ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_bitset(c1), const_CAST_run(c2), &result) ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1929| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1930| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1931| 0| return result; 1932| | 1933| 0| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (1933:9): [True: 0, False: 5.81k] ------------------ 1934| 0| *result_type = run_bitset_container_iandnot( ------------------ | Branch (1934:28): [True: 0, False: 0] ------------------ 1935| 0| CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_bitset(c2), &result) ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1936| 0| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 1937| 0| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1938| 0| return result; 1939| | 1940| 50| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 50| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 50|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 50|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (1940:9): [True: 50, False: 5.76k] ------------------ 1941| 50| *result_type = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 50|#define ARRAY_CONTAINER_TYPE 2 ------------------ 1942| 50| array_run_container_iandnot(CAST_array(c1), const_CAST_run(c2)); ------------------ | | 54| 50|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 50|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_run_container_iandnot(CAST_array(c1), const_CAST_run(c2)); ------------------ | | 78| 50|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 50|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1943| 50| return c1; 1944| | 1945| 742| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 742| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 742|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 742|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (1945:9): [True: 742, False: 5.07k] ------------------ 1946| 742| *result_type = (uint8_t)run_array_container_iandnot( 1947| 742| CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 77| 742|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 742|#define CAST(type, value) ((type)value) | | ------------------ ------------------ CAST_run(c1), const_CAST_array(c2), &result); ------------------ | | 55| 742|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 742|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1948| 742| return result; 1949| | 1950| 0| default: ------------------ | Branch (1950:9): [True: 0, False: 5.81k] ------------------ 1951| 0| assert(false); 1952| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1953| 0| return NULL; 1954| 5.81k| } 1955| 5.81k|} roaring.c:container_rank: 2209| 6.24k| uint16_t x) { 2210| 6.24k| c = container_unwrap_shared(c, &type); 2211| 6.24k| switch (type) { 2212| 0| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2212:9): [True: 0, False: 6.24k] ------------------ 2213| 0| return bitset_container_rank(const_CAST_bitset(c), x); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2214| 3.70k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.70k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2214:9): [True: 3.70k, False: 2.54k] ------------------ 2215| 3.70k| return array_container_rank(const_CAST_array(c), x); ------------------ | | 55| 3.70k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.70k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2216| 2.54k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 2.54k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2216:9): [True: 2.54k, False: 3.70k] ------------------ 2217| 2.54k| return run_container_rank(const_CAST_run(c), x); ------------------ | | 78| 2.54k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2218| 0| default: ------------------ | Branch (2218:9): [True: 0, False: 6.24k] ------------------ 2219| 0| assert(false); 2220| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2221| 6.24k| } 2222| 6.24k| assert(false); 2223| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2224| 0| return false; 2225| 6.24k|} roaring.c:container_shrink_to_fit: 279| 6.24k|static inline int container_shrink_to_fit(container_t *c, uint8_t type) { 280| 6.24k| c = container_mutable_unwrap_shared(c, &type); 281| 6.24k| switch (type) { ------------------ | Branch (281:13): [True: 6.24k, False: 0] ------------------ 282| 0| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (282:9): [True: 0, False: 6.24k] ------------------ 283| 0| return 0; // no shrinking possible 284| 3.70k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.70k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (284:9): [True: 3.70k, False: 2.54k] ------------------ 285| 3.70k| return array_container_shrink_to_fit(CAST_array(c)); ------------------ | | 54| 3.70k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 3.70k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 286| 2.54k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 2.54k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (286:9): [True: 2.54k, False: 3.70k] ------------------ 287| 2.54k| return run_container_shrink_to_fit(CAST_run(c)); ------------------ | | 77| 2.54k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 288| 6.24k| } 289| 6.24k| assert(false); 290| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 291| 0| return 0; // unreached 292| 6.24k|} roaring.c:container_iterate: 1964| 63.3k| void *ptr) { 1965| 63.3k| c = container_unwrap_shared(c, &type); 1966| 63.3k| switch (type) { 1967| 1.56k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 1.56k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (1967:9): [True: 1.56k, False: 61.8k] ------------------ 1968| 1.56k| return bitset_container_iterate(const_CAST_bitset(c), base, ------------------ | | 53| 1.56k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.56k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1969| 1.56k| iterator, ptr); 1970| 6.22k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 6.22k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (1970:9): [True: 6.22k, False: 57.1k] ------------------ 1971| 6.22k| return array_container_iterate(const_CAST_array(c), base, iterator, ------------------ | | 55| 6.22k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 6.22k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1972| 6.22k| ptr); 1973| 55.6k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 55.6k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (1973:9): [True: 55.6k, False: 7.79k] ------------------ 1974| 55.6k| return run_container_iterate(const_CAST_run(c), base, iterator, ------------------ | | 78| 55.6k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 55.6k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1975| 55.6k| ptr); 1976| 0| default: ------------------ | Branch (1976:9): [True: 0, False: 63.3k] ------------------ 1977| 0| assert(false); 1978| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1979| 63.3k| } 1980| 63.3k| assert(false); 1981| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 1982| 0| return false; 1983| 63.3k|} roaring.c:container_maximum: 2171| 12.3k|static inline uint16_t container_maximum(const container_t *c, uint8_t type) { 2172| 12.3k| c = container_unwrap_shared(c, &type); 2173| 12.3k| switch (type) { 2174| 2.43k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 2.43k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2174:9): [True: 2.43k, False: 9.94k] ------------------ 2175| 2.43k| return bitset_container_maximum(const_CAST_bitset(c)); ------------------ | | 53| 2.43k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 2.43k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2176| 8.59k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 8.59k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2176:9): [True: 8.59k, False: 3.79k] ------------------ 2177| 8.59k| return array_container_maximum(const_CAST_array(c)); ------------------ | | 55| 8.59k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 8.59k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2178| 1.35k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 1.35k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2178:9): [True: 1.35k, False: 11.0k] ------------------ 2179| 1.35k| return run_container_maximum(const_CAST_run(c)); ------------------ | | 78| 1.35k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.35k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2180| 0| default: ------------------ | Branch (2180:9): [True: 0, False: 12.3k] ------------------ 2181| 0| assert(false); 2182| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2183| 12.3k| } 2184| 12.3k| assert(false); 2185| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2186| 0| return false; 2187| 12.3k|} roaring.c:container_equals: 620| 77.7k| const container_t *c2, uint8_t type2) { 621| 77.7k| c1 = container_unwrap_shared(c1, &type1); 622| 77.7k| c2 = container_unwrap_shared(c2, &type2); 623| 77.7k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 77.7k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 624| 66.9k| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 66.9k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 66.9k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 66.9k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (624:9): [True: 66.9k, False: 10.7k] ------------------ 625| 66.9k| return bitset_container_equals(const_CAST_bitset(c1), ------------------ | | 53| 66.9k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 66.9k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 626| 66.9k| const_CAST_bitset(c2)); ------------------ | | 53| 66.9k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 66.9k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 627| | 628| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (628:9): [True: 0, False: 77.7k] ------------------ 629| 0| return run_container_equals_bitset(const_CAST_run(c2), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 630| 0| const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 631| | 632| 0| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (632:9): [True: 0, False: 77.7k] ------------------ 633| 0| return run_container_equals_bitset(const_CAST_run(c1), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 634| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 635| | 636| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (636:9): [True: 0, False: 77.7k] ------------------ 637| | // java would always return false? 638| 0| return array_container_equal_bitset(const_CAST_array(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 639| 0| const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 640| | 641| 457| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 457| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 457|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 457|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (641:9): [True: 457, False: 77.2k] ------------------ 642| | // java would always return false? 643| 457| return array_container_equal_bitset(const_CAST_array(c1), ------------------ | | 55| 457|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 457|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 644| 457| const_CAST_bitset(c2)); ------------------ | | 53| 457|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 457|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 645| | 646| 641| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 641| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 641|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 641|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (646:9): [True: 641, False: 77.0k] ------------------ 647| 641| return run_container_equals_array(const_CAST_run(c2), ------------------ | | 78| 641|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 641|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 648| 641| const_CAST_array(c1)); ------------------ | | 55| 641|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 641|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 649| | 650| 256| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 256| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 256|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 256|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (650:9): [True: 256, False: 77.4k] ------------------ 651| 256| return run_container_equals_array(const_CAST_run(c1), ------------------ | | 78| 256|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 256|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 652| 256| const_CAST_array(c2)); ------------------ | | 55| 256|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 256|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 653| | 654| 9.23k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 9.23k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 9.23k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 9.23k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (654:9): [True: 9.23k, False: 68.4k] ------------------ 655| 9.23k| return array_container_equals(const_CAST_array(c1), ------------------ | | 55| 9.23k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 9.23k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 656| 9.23k| const_CAST_array(c2)); ------------------ | | 55| 9.23k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 9.23k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 657| | 658| 133| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 133| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 133|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 133|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (658:9): [True: 133, False: 77.5k] ------------------ 659| 133| return run_container_equals(const_CAST_run(c1), const_CAST_run(c2)); ------------------ | | 78| 133|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 133|#define CAST(type, value) ((type)value) | | ------------------ ------------------ return run_container_equals(const_CAST_run(c1), const_CAST_run(c2)); ------------------ | | 78| 133|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 133|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 660| | 661| 0| default: ------------------ | Branch (661:9): [True: 0, False: 77.7k] ------------------ 662| 0| assert(false); 663| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 664| 0| return false; 665| 77.7k| } 666| 77.7k|} roaring.c:container_is_subset: 673| 10.7k| const container_t *c2, uint8_t type2) { 674| 10.7k| c1 = container_unwrap_shared(c1, &type1); 675| 10.7k| c2 = container_unwrap_shared(c2, &type2); 676| 10.7k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 10.7k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 677| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (677:9): [True: 0, False: 10.7k] ------------------ 678| 0| return bitset_container_is_subset(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 679| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 680| | 681| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (681:9): [True: 0, False: 10.7k] ------------------ 682| 0| return bitset_container_is_subset_run(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 683| 0| const_CAST_run(c2)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 684| | 685| 0| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (685:9): [True: 0, False: 10.7k] ------------------ 686| 0| return run_container_is_subset_bitset(const_CAST_run(c1), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 687| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 688| | 689| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (689:9): [True: 0, False: 10.7k] ------------------ 690| 0| return false; // by construction, size(c1) > size(c2) 691| | 692| 2.20k| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 2.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 2.20k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (692:9): [True: 2.20k, False: 8.52k] ------------------ 693| 2.20k| return array_container_is_subset_bitset(const_CAST_array(c1), ------------------ | | 55| 2.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 694| 2.20k| const_CAST_bitset(c2)); ------------------ | | 53| 2.20k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 2.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 695| | 696| 2.27k| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 2.27k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.27k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 2.27k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (696:9): [True: 2.27k, False: 8.44k] ------------------ 697| 2.27k| return array_container_is_subset_run(const_CAST_array(c1), ------------------ | | 55| 2.27k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.27k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 698| 2.27k| const_CAST_run(c2)); ------------------ | | 78| 2.27k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.27k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 699| | 700| 391| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 391| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 391|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 391|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (700:9): [True: 391, False: 10.3k] ------------------ 701| 391| return run_container_is_subset_array(const_CAST_run(c1), ------------------ | | 78| 391|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 391|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 702| 391| const_CAST_array(c2)); ------------------ | | 55| 391|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 391|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 703| | 704| 5.59k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 5.59k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 5.59k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 5.59k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (704:9): [True: 5.59k, False: 5.12k] ------------------ 705| 5.59k| return array_container_is_subset(const_CAST_array(c1), ------------------ | | 55| 5.59k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 5.59k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 706| 5.59k| const_CAST_array(c2)); ------------------ | | 55| 5.59k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 5.59k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 707| | 708| 256| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 256| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 256|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 256|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (708:9): [True: 256, False: 10.4k] ------------------ 709| 256| return run_container_is_subset(const_CAST_run(c1), ------------------ | | 78| 256|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 256|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 710| 256| const_CAST_run(c2)); ------------------ | | 78| 256|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 256|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 711| | 712| 0| default: ------------------ | Branch (712:9): [True: 0, False: 10.7k] ------------------ 713| 0| assert(false); 714| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 715| 0| return false; 716| 10.7k| } 717| 10.7k|} roaring.c:container_inot_range: 2110| 6.45k| uint8_t *result_type) { 2111| 6.45k| c = get_writable_copy_if_shared(c, &type); 2112| 6.45k| container_t *result = NULL; 2113| 6.45k| switch (type) { 2114| 1.23k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 1.23k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2114:9): [True: 1.23k, False: 5.21k] ------------------ 2115| 1.23k| *result_type = bitset_container_negation_range_inplace( ------------------ | Branch (2115:28): [True: 1.20k, False: 33] ------------------ 2116| 1.23k| CAST_bitset(c), range_start, range_end, &result) ------------------ | | 52| 1.23k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.23k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2117| 1.23k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 1.20k|#define BITSET_CONTAINER_TYPE 1 ------------------ 2118| 1.23k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.27k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2119| 1.23k| return result; 2120| 3.49k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.49k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2120:9): [True: 3.49k, False: 2.95k] ------------------ 2121| 3.49k| *result_type = array_container_negation_range_inplace( ------------------ | Branch (2121:28): [True: 130, False: 3.36k] ------------------ 2122| 3.49k| CAST_array(c), range_start, range_end, &result) ------------------ | | 54| 3.49k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 3.49k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2123| 3.49k| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 130|#define BITSET_CONTAINER_TYPE 1 ------------------ 2124| 3.49k| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 6.86k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2125| 3.49k| return result; 2126| 1.71k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 1.71k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2126:9): [True: 1.71k, False: 4.73k] ------------------ 2127| 1.71k| *result_type = (uint8_t)run_container_negation_range_inplace( 2128| 1.71k| CAST_run(c), range_start, range_end, &result); ------------------ | | 77| 1.71k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.71k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2129| 1.71k| return result; 2130| | 2131| 0| default: ------------------ | Branch (2131:9): [True: 0, False: 6.45k] ------------------ 2132| 0| assert(false); 2133| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2134| 6.45k| } 2135| 6.45k| assert(false); 2136| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2137| 0| return NULL; 2138| 6.45k|} roaring.c:container_inot: 2076| 6.84k| uint8_t *result_type) { 2077| 6.84k| c = get_writable_copy_if_shared(c, &type); 2078| 6.84k| container_t *result = NULL; 2079| 6.84k| switch (type) { 2080| 61| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 61|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2080:9): [True: 61, False: 6.78k] ------------------ 2081| 61| *result_type = 2082| 61| bitset_container_negation_inplace(CAST_bitset(c), &result) ------------------ | | 52| 61|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 61|#define CAST(type, value) ((type)value) | | ------------------ ------------------ | Branch (2082:17): [True: 23, False: 38] ------------------ 2083| 61| ? BITSET_CONTAINER_TYPE ------------------ | | 48| 23|#define BITSET_CONTAINER_TYPE 1 ------------------ 2084| 61| : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 99|#define ARRAY_CONTAINER_TYPE 2 ------------------ 2085| 61| return result; 2086| 155| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 155|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2086:9): [True: 155, False: 6.69k] ------------------ 2087| | // will never be inplace 2088| 155| result = bitset_container_create(); 2089| 155| *result_type = BITSET_CONTAINER_TYPE; ------------------ | | 48| 155|#define BITSET_CONTAINER_TYPE 1 ------------------ 2090| 155| array_container_negation(CAST_array(c), CAST_bitset(result)); ------------------ | | 54| 155|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 155|#define CAST(type, value) ((type)value) | | ------------------ ------------------ array_container_negation(CAST_array(c), CAST_bitset(result)); ------------------ | | 52| 155|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 155|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2091| 155| array_container_free(CAST_array(c)); ------------------ | | 54| 155|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 155|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2092| 155| return result; 2093| 6.63k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 6.63k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2093:9): [True: 6.63k, False: 216] ------------------ 2094| 6.63k| *result_type = 2095| 6.63k| (uint8_t)run_container_negation_inplace(CAST_run(c), &result); ------------------ | | 77| 6.63k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 6.63k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2096| 6.63k| return result; 2097| | 2098| 0| default: ------------------ | Branch (2098:9): [True: 0, False: 6.84k] ------------------ 2099| 0| assert(false); 2100| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2101| 6.84k| } 2102| 6.84k| assert(false); 2103| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2104| 0| return NULL; 2105| 6.84k|} roaring.c:container_unwrap_shared: 106| 1.08M| const container_t *candidate_shared_container, uint8_t *type) { 107| 1.08M| if (*type == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 1.08M|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (107:9): [True: 0, False: 1.08M] ------------------ 108| 0| *type = const_CAST_shared(candidate_shared_container)->typecode; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 109| 0| assert(*type != SHARED_CONTAINER_TYPE); 110| 0| return const_CAST_shared(candidate_shared_container)->container; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 111| 1.08M| } else { 112| 1.08M| return candidate_shared_container; 113| 1.08M| } 114| 1.08M|} roaring.c:container_mutable_unwrap_shared: 118| 6.24k| uint8_t *type) { 119| 6.24k| if (*type == SHARED_CONTAINER_TYPE) { // the passed in container is shared ------------------ | | 51| 6.24k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (119:9): [True: 0, False: 6.24k] ------------------ 120| 0| *type = CAST_shared(c)->typecode; ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 121| 0| assert(*type != SHARED_CONTAINER_TYPE); 122| 0| return CAST_shared(c)->container; // return the enclosed container ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 123| 6.24k| } else { 124| 6.24k| return c; // wasn't shared, so return as-is 125| 6.24k| } 126| 6.24k|} roaring.c:container_minimum: 2189| 6.19k|static inline uint16_t container_minimum(const container_t *c, uint8_t type) { 2190| 6.19k| c = container_unwrap_shared(c, &type); 2191| 6.19k| switch (type) { 2192| 1.15k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 1.15k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2192:9): [True: 1.15k, False: 5.04k] ------------------ 2193| 1.15k| return bitset_container_minimum(const_CAST_bitset(c)); ------------------ | | 53| 1.15k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.15k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2194| 3.54k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.54k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2194:9): [True: 3.54k, False: 2.64k] ------------------ 2195| 3.54k| return array_container_minimum(const_CAST_array(c)); ------------------ | | 55| 3.54k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2196| 1.49k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 1.49k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2196:9): [True: 1.49k, False: 4.69k] ------------------ 2197| 1.49k| return run_container_minimum(const_CAST_run(c)); ------------------ | | 78| 1.49k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.49k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2198| 0| default: ------------------ | Branch (2198:9): [True: 0, False: 6.19k] ------------------ 2199| 0| assert(false); 2200| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2201| 6.19k| } 2202| 6.19k| assert(false); 2203| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2204| 0| return false; 2205| 6.19k|} roaring.c:container_select: 2150| 6.24k| uint32_t *element) { 2151| 6.24k| c = container_unwrap_shared(c, &type); 2152| 6.24k| switch (type) { 2153| 0| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2153:9): [True: 0, False: 6.24k] ------------------ 2154| 0| return bitset_container_select(const_CAST_bitset(c), start_rank, ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2155| 0| rank, element); 2156| 3.70k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 3.70k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2156:9): [True: 3.70k, False: 2.54k] ------------------ 2157| 3.70k| return array_container_select(const_CAST_array(c), start_rank, rank, ------------------ | | 55| 3.70k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.70k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2158| 3.70k| element); 2159| 2.54k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 2.54k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2159:9): [True: 2.54k, False: 3.70k] ------------------ 2160| 2.54k| return run_container_select(const_CAST_run(c), start_rank, rank, ------------------ | | 78| 2.54k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2161| 2.54k| element); 2162| 0| default: ------------------ | Branch (2162:9): [True: 0, False: 6.24k] ------------------ 2163| 0| assert(false); 2164| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2165| 6.24k| } 2166| 6.24k| assert(false); 2167| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2168| 0| return false; 2169| 6.24k|} roaring.c:container_intersect: 865| 6.12k| const container_t *c2, uint8_t type2) { 866| 6.12k| c1 = container_unwrap_shared(c1, &type1); 867| 6.12k| c2 = container_unwrap_shared(c2, &type2); 868| 6.12k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 6.12k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 869| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (869:9): [True: 0, False: 6.12k] ------------------ 870| 0| return bitset_container_intersect(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 871| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 872| | 873| 2.30k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 2.30k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.30k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (873:9): [True: 2.30k, False: 3.82k] ------------------ 874| 2.30k| return array_container_intersect(const_CAST_array(c1), ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 875| 2.30k| const_CAST_array(c2)); ------------------ | | 55| 2.30k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.30k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 876| | 877| 690| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 690| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 690|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (877:9): [True: 690, False: 5.43k] ------------------ 878| 690| return run_container_intersect(const_CAST_run(c1), ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 879| 690| const_CAST_run(c2)); ------------------ | | 78| 690|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 690|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 880| | 881| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (881:9): [True: 0, False: 6.12k] ------------------ 882| 0| return array_bitset_container_intersect(const_CAST_array(c2), ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 883| 0| const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 884| | 885| 531| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 531| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 531|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 531|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (885:9): [True: 531, False: 5.59k] ------------------ 886| 531| return array_bitset_container_intersect(const_CAST_array(c1), ------------------ | | 55| 531|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 887| 531| const_CAST_bitset(c2)); ------------------ | | 53| 531|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 531|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 888| | 889| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (889:9): [True: 0, False: 6.12k] ------------------ 890| 0| return run_bitset_container_intersect(const_CAST_run(c2), ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 891| 0| const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 892| | 893| 613| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 613| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 613|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 613|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (893:9): [True: 613, False: 5.51k] ------------------ 894| 613| return run_bitset_container_intersect(const_CAST_run(c1), ------------------ | | 78| 613|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 895| 613| const_CAST_bitset(c2)); ------------------ | | 53| 613|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 613|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 896| | 897| 783| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 783| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 783|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 783|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (897:9): [True: 783, False: 5.34k] ------------------ 898| 783| return array_run_container_intersect(const_CAST_array(c1), ------------------ | | 55| 783|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 899| 783| const_CAST_run(c2)); ------------------ | | 78| 783|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 783|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 900| | 901| 1.20k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 1.20k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 1.20k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 1.20k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (901:9): [True: 1.20k, False: 4.92k] ------------------ 902| 1.20k| return array_run_container_intersect(const_CAST_array(c2), ------------------ | | 55| 1.20k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 903| 1.20k| const_CAST_run(c1)); ------------------ | | 78| 1.20k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 1.20k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 904| | 905| 0| default: ------------------ | Branch (905:9): [True: 0, False: 6.12k] ------------------ 906| 0| assert(false); 907| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 908| 0| return 0; 909| 6.12k| } 910| 6.12k|} roaring.c:container_and_cardinality: 814| 24.5k| uint8_t type2) { 815| 24.5k| c1 = container_unwrap_shared(c1, &type1); 816| 24.5k| c2 = container_unwrap_shared(c2, &type2); 817| 24.5k| switch (PAIR_CONTAINER_TYPES(type1, type2)) { ------------------ | | 62| 24.5k|#define PAIR_CONTAINER_TYPES(type1, type2) (4 * (type1) + (type2)) ------------------ 818| 0| case CONTAINER_PAIR(BITSET, BITSET): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (818:9): [True: 0, False: 24.5k] ------------------ 819| 0| return bitset_container_and_justcard(const_CAST_bitset(c1), ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 820| 0| const_CAST_bitset(c2)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 821| | 822| 9.21k| case CONTAINER_PAIR(ARRAY, ARRAY): ------------------ | | 65| 9.21k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 9.21k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 9.21k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (822:9): [True: 9.21k, False: 15.2k] ------------------ 823| 9.21k| return array_container_intersection_cardinality( 824| 9.21k| const_CAST_array(c1), const_CAST_array(c2)); ------------------ | | 55| 9.21k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 9.21k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_array(c2)); ------------------ | | 55| 9.21k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 9.21k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 825| | 826| 2.76k| case CONTAINER_PAIR(RUN, RUN): ------------------ | | 65| 2.76k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 2.76k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 2.76k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (826:9): [True: 2.76k, False: 21.7k] ------------------ 827| 2.76k| return run_container_intersection_cardinality(const_CAST_run(c1), ------------------ | | 78| 2.76k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.76k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 828| 2.76k| const_CAST_run(c2)); ------------------ | | 78| 2.76k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.76k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 829| | 830| 0| case CONTAINER_PAIR(BITSET, ARRAY): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 0|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (830:9): [True: 0, False: 24.5k] ------------------ 831| 0| return array_bitset_container_intersection_cardinality( 832| 0| const_CAST_array(c2), const_CAST_bitset(c1)); ------------------ | | 55| 0|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 833| | 834| 2.12k| case CONTAINER_PAIR(ARRAY, BITSET): ------------------ | | 65| 2.12k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 2.12k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 2.12k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (834:9): [True: 2.12k, False: 22.3k] ------------------ 835| 2.12k| return array_bitset_container_intersection_cardinality( 836| 2.12k| const_CAST_array(c1), const_CAST_bitset(c2)); ------------------ | | 55| 2.12k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 2.12k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_bitset(c2)); ------------------ | | 53| 2.12k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 2.12k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 837| | 838| 0| case CONTAINER_PAIR(BITSET, RUN): ------------------ | | 65| 0| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 0|#define BITSET_CONTAINER_TYPE 1 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 0|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (838:9): [True: 0, False: 24.5k] ------------------ 839| 0| return run_bitset_container_intersection_cardinality( 840| 0| const_CAST_run(c2), const_CAST_bitset(c1)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c2), const_CAST_bitset(c1)); ------------------ | | 53| 0|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 841| | 842| 2.45k| case CONTAINER_PAIR(RUN, BITSET): ------------------ | | 65| 2.45k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 2.45k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 48| 2.45k|#define BITSET_CONTAINER_TYPE 1 | | ------------------ ------------------ | Branch (842:9): [True: 2.45k, False: 22.0k] ------------------ 843| 2.45k| return run_bitset_container_intersection_cardinality( 844| 2.45k| const_CAST_run(c1), const_CAST_bitset(c2)); ------------------ | | 78| 2.45k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.45k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_run(c1), const_CAST_bitset(c2)); ------------------ | | 53| 2.45k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 2.45k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 845| | 846| 3.13k| case CONTAINER_PAIR(ARRAY, RUN): ------------------ | | 65| 3.13k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 3.13k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 3.13k|#define RUN_CONTAINER_TYPE 3 | | ------------------ ------------------ | Branch (846:9): [True: 3.13k, False: 21.3k] ------------------ 847| 3.13k| return array_run_container_intersection_cardinality( 848| 3.13k| const_CAST_array(c1), const_CAST_run(c2)); ------------------ | | 55| 3.13k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 3.13k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c1), const_CAST_run(c2)); ------------------ | | 78| 3.13k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 3.13k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 849| | 850| 4.82k| case CONTAINER_PAIR(RUN, ARRAY): ------------------ | | 65| 4.82k| (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 50| 4.82k|#define RUN_CONTAINER_TYPE 3 | | ------------------ | | (4 * (name1##_CONTAINER_TYPE) + (name2##_CONTAINER_TYPE)) | | ------------------ | | | | 49| 4.82k|#define ARRAY_CONTAINER_TYPE 2 | | ------------------ ------------------ | Branch (850:9): [True: 4.82k, False: 19.6k] ------------------ 851| 4.82k| return array_run_container_intersection_cardinality( 852| 4.82k| const_CAST_array(c2), const_CAST_run(c1)); ------------------ | | 55| 4.82k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 4.82k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ const_CAST_array(c2), const_CAST_run(c1)); ------------------ | | 78| 4.82k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 4.82k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 853| | 854| 0| default: ------------------ | Branch (854:9): [True: 0, False: 24.5k] ------------------ 855| 0| assert(false); 856| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 857| 0| return 0; 858| 24.5k| } 859| 24.5k|} roaring.c:container_contains_range: 596| 493|) { 597| 493| c = container_unwrap_shared(c, &typecode); 598| 493| switch (typecode) { 599| 219| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 219|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (599:9): [True: 219, False: 274] ------------------ 600| 219| return bitset_container_get_range(const_CAST_bitset(c), range_start, ------------------ | | 53| 219|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 219|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 601| 219| range_end); 602| 83| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 83|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (602:9): [True: 83, False: 410] ------------------ 603| 83| return array_container_contains_range(const_CAST_array(c), ------------------ | | 55| 83|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 83|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 604| 83| range_start, range_end); 605| 191| case RUN_CONTAINER_TYPE: ------------------ | | 50| 191|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (605:9): [True: 191, False: 302] ------------------ 606| 191| return run_container_contains_range(const_CAST_run(c), range_start, ------------------ | | 78| 191|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 191|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 607| 191| range_end); 608| 0| default: ------------------ | Branch (608:9): [True: 0, False: 493] ------------------ 609| 0| assert(false); 610| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 611| 0| return false; 612| 493| } 613| 493|} roaring_array.c:container_to_uint32_array: 465| 63.2k| uint8_t typecode, uint32_t base) { 466| 63.2k| c = container_unwrap_shared(c, &typecode); 467| 63.2k| switch (typecode) { ------------------ | Branch (467:13): [True: 63.2k, False: 0] ------------------ 468| 1.40k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 1.40k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (468:9): [True: 1.40k, False: 61.8k] ------------------ 469| 1.40k| return bitset_container_to_uint32_array(output, 470| 1.40k| const_CAST_bitset(c), base); ------------------ | | 53| 1.40k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.40k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 471| 5.62k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 5.62k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (471:9): [True: 5.62k, False: 57.6k] ------------------ 472| 5.62k| return array_container_to_uint32_array(output, const_CAST_array(c), ------------------ | | 55| 5.62k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 5.62k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 473| 5.62k| base); 474| 56.2k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 56.2k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (474:9): [True: 56.2k, False: 7.03k] ------------------ 475| 56.2k| return run_container_to_uint32_array(output, const_CAST_run(c), ------------------ | | 78| 56.2k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 56.2k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 476| 56.2k| base); 477| 63.2k| } 478| 63.2k| assert(false); 479| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 480| 0| return 0; // unreached 481| 63.2k|} roaring_array.c:container_unwrap_shared: 106| 364k| const container_t *candidate_shared_container, uint8_t *type) { 107| 364k| if (*type == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 364k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (107:9): [True: 0, False: 364k] ------------------ 108| 0| *type = const_CAST_shared(candidate_shared_container)->typecode; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 109| 0| assert(*type != SHARED_CONTAINER_TYPE); 110| 0| return const_CAST_shared(candidate_shared_container)->container; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 111| 364k| } else { 112| 364k| return candidate_shared_container; 113| 364k| } 114| 364k|} roaring_array.c:get_container_type: 129| 153k|static inline uint8_t get_container_type(const container_t *c, uint8_t type) { 130| 153k| if (type == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 153k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (130:9): [True: 0, False: 153k] ------------------ 131| 0| return const_CAST_shared(c)->typecode; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 132| 153k| } else { 133| 153k| return type; 134| 153k| } 135| 153k|} roaring_array.c:container_size_in_bytes: 403| 153k| uint8_t typecode) { 404| 153k| c = container_unwrap_shared(c, &typecode); 405| 153k| switch (typecode) { ------------------ | Branch (405:13): [True: 153k, False: 0] ------------------ 406| 133k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 133k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (406:9): [True: 133k, False: 19.8k] ------------------ 407| 133k| return bitset_container_size_in_bytes(const_CAST_bitset(c)); ------------------ | | 53| 133k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 133k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 408| 17.3k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 17.3k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (408:9): [True: 17.3k, False: 136k] ------------------ 409| 17.3k| return array_container_size_in_bytes(const_CAST_array(c)); ------------------ | | 55| 17.3k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 17.3k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 410| 2.54k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 2.54k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (410:9): [True: 2.54k, False: 151k] ------------------ 411| 2.54k| return run_container_size_in_bytes(const_CAST_run(c)); ------------------ | | 78| 2.54k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 2.54k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 412| 153k| } 413| 153k| assert(false); 414| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 415| 0| return 0; // unreached 416| 153k|} roaring_array.c:container_get_cardinality: 244| 73.8k| uint8_t typecode) { 245| 73.8k| c = container_unwrap_shared(c, &typecode); 246| 73.8k| switch (typecode) { ------------------ | Branch (246:13): [True: 73.8k, False: 0] ------------------ 247| 66.9k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 66.9k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (247:9): [True: 66.9k, False: 6.82k] ------------------ 248| 66.9k| return bitset_container_cardinality(const_CAST_bitset(c)); ------------------ | | 53| 66.9k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 66.9k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 249| 6.82k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 6.82k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (249:9): [True: 6.82k, False: 66.9k] ------------------ 250| 6.82k| return array_container_cardinality(const_CAST_array(c)); ------------------ | | 55| 6.82k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 6.82k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 251| 0| case RUN_CONTAINER_TYPE: ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (251:9): [True: 0, False: 73.8k] ------------------ 252| 0| return run_container_cardinality(const_CAST_run(c)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 253| 73.8k| } 254| 73.8k| assert(false); 255| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 256| 0| return 0; // unreached 257| 73.8k|} roaring_array.c:container_write: 382| 73.8k| char *buf) { 383| 73.8k| c = container_unwrap_shared(c, &typecode); 384| 73.8k| switch (typecode) { ------------------ | Branch (384:13): [True: 73.8k, False: 0] ------------------ 385| 66.9k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 66.9k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (385:9): [True: 66.9k, False: 6.82k] ------------------ 386| 66.9k| return bitset_container_write(const_CAST_bitset(c), buf); ------------------ | | 53| 66.9k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 66.9k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 387| 6.82k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 6.82k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (387:9): [True: 6.82k, False: 66.9k] ------------------ 388| 6.82k| return array_container_write(const_CAST_array(c), buf); ------------------ | | 55| 6.82k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 6.82k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 389| 0| case RUN_CONTAINER_TYPE: ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (389:9): [True: 0, False: 73.8k] ------------------ 390| 0| return run_container_write(const_CAST_run(c), buf); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 391| 73.8k| } 392| 73.8k| assert(false); 393| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 394| 0| return 0; // unreached 395| 73.8k|} containers.c:container_unwrap_shared: 106| 6.40k| const container_t *candidate_shared_container, uint8_t *type) { 107| 6.40k| if (*type == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 6.40k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (107:9): [True: 0, False: 6.40k] ------------------ 108| 0| *type = const_CAST_shared(candidate_shared_container)->typecode; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 109| 0| assert(*type != SHARED_CONTAINER_TYPE); 110| 0| return const_CAST_shared(candidate_shared_container)->container; ------------------ | | 80| 0|#define const_CAST_shared(c) CAST(const shared_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 111| 6.40k| } else { 112| 6.40k| return candidate_shared_container; 113| 6.40k| } 114| 6.40k|} containers.c:container_maximum: 2171| 6.17k|static inline uint16_t container_maximum(const container_t *c, uint8_t type) { 2172| 6.17k| c = container_unwrap_shared(c, &type); 2173| 6.17k| switch (type) { 2174| 1.79k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 1.79k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (2174:9): [True: 1.79k, False: 4.37k] ------------------ 2175| 1.79k| return bitset_container_maximum(const_CAST_bitset(c)); ------------------ | | 53| 1.79k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.79k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2176| 4.37k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 4.37k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (2176:9): [True: 4.37k, False: 1.79k] ------------------ 2177| 4.37k| return array_container_maximum(const_CAST_array(c)); ------------------ | | 55| 4.37k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 4.37k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2178| 0| case RUN_CONTAINER_TYPE: ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (2178:9): [True: 0, False: 6.17k] ------------------ 2179| 0| return run_container_maximum(const_CAST_run(c)); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 2180| 0| default: ------------------ | Branch (2180:9): [True: 0, False: 6.17k] ------------------ 2181| 0| assert(false); 2182| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2183| 6.17k| } 2184| 6.17k| assert(false); 2185| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 2186| 0| return false; 2187| 6.17k|} run_container_contains: 260| 9.61k|inline bool run_container_contains(const run_container_t *run, uint16_t pos) { 261| 9.61k| int32_t index = interleavedBinarySearch(run->runs, run->n_runs, pos); 262| 9.61k| if (index >= 0) return true; ------------------ | Branch (262:9): [True: 6.96k, False: 2.64k] ------------------ 263| 2.64k| index = -index - 2; // points to preceding value, possibly -1 264| 2.64k| if (index != -1) { // possible match ------------------ | Branch (264:9): [True: 908, False: 1.74k] ------------------ 265| 908| int32_t offset = pos - run->runs[index].value; 266| 908| int32_t le = run->runs[index].length; 267| 908| if (offset <= le) return true; ------------------ | Branch (267:13): [True: 597, False: 311] ------------------ 268| 908| } 269| 2.05k| return false; 270| 2.64k|} interleavedBinarySearch: 112| 425k| uint16_t ikey) { 113| 425k| int32_t low = 0; 114| 425k| int32_t high = lenarray - 1; 115| 1.97M| while (low <= high) { ------------------ | Branch (115:12): [True: 1.73M, False: 237k] ------------------ 116| 1.73M| int32_t middleIndex = (low + high) >> 1; 117| 1.73M| uint16_t middleValue = array[middleIndex].value; 118| 1.73M| if (middleValue < ikey) { ------------------ | Branch (118:13): [True: 683k, False: 1.05M] ------------------ 119| 683k| low = middleIndex + 1; 120| 1.05M| } else if (middleValue > ikey) { ------------------ | Branch (120:20): [True: 863k, False: 187k] ------------------ 121| 863k| high = middleIndex - 1; 122| 863k| } else { 123| 187k| return middleIndex; 124| 187k| } 125| 1.73M| } 126| 237k| return -(low + 1); 127| 425k|} run_container_minimum: 563| 17.2k|inline uint16_t run_container_minimum(const run_container_t *run) { 564| 17.2k| if (run->n_runs == 0) return 0; ------------------ | Branch (564:9): [True: 0, False: 17.2k] ------------------ 565| 17.2k| return run->runs[0].value; 566| 17.2k|} run_container_maximum: 569| 16.7k|inline uint16_t run_container_maximum(const run_container_t *run) { 570| 16.7k| if (run->n_runs == 0) return 0; ------------------ | Branch (570:9): [True: 0, False: 16.7k] ------------------ 571| 16.7k| return run->runs[run->n_runs - 1].value + run->runs[run->n_runs - 1].length; 572| 16.7k|} roaring.c:run_container_create_range: 532| 85.6k| uint32_t stop) { 533| 85.6k| run_container_t *rc = run_container_create_given_capacity(1); 534| 85.6k| if (rc) { ------------------ | Branch (534:9): [True: 85.6k, False: 0] ------------------ 535| 85.6k| rle16_t r; 536| 85.6k| r.value = (uint16_t)start; 537| 85.6k| r.length = (uint16_t)(stop - start - 1); 538| 85.6k| run_container_append_first(rc, r); 539| 85.6k| } 540| 85.6k| return rc; 541| 85.6k|} roaring.c:run_container_append_first: 351| 85.6k| rle16_t vl) { 352| 85.6k| run->runs[run->n_runs] = vl; 353| 85.6k| run->n_runs++; 354| 85.6k| return vl; 355| 85.6k|} roaring.c:rle16_count_greater: 178| 476| int32_t lenarray, uint16_t key) { 179| 476| if (lenarray == 0) return 0; ------------------ | Branch (179:9): [True: 0, False: 476] ------------------ 180| 476| int32_t low = 0; 181| 476| int32_t high = lenarray - 1; 182| 3.61k| while (low <= high) { ------------------ | Branch (182:12): [True: 3.17k, False: 435] ------------------ 183| 3.17k| int32_t middleIndex = (low + high) >> 1; 184| 3.17k| uint16_t min_value = array[middleIndex].value; 185| 3.17k| uint16_t max_value = 186| 3.17k| array[middleIndex].value + array[middleIndex].length; 187| 3.17k| if (max_value < key) { ------------------ | Branch (187:13): [True: 2.90k, False: 276] ------------------ 188| 2.90k| low = middleIndex + 1; 189| 2.90k| } else if (key + UINT32_C(1) < min_value) { // uint32 arithmetic ------------------ | Branch (189:20): [True: 235, False: 41] ------------------ 190| 235| high = middleIndex - 1; 191| 235| } else { 192| 41| return lenarray - (middleIndex + 1); 193| 41| } 194| 3.17k| } 195| 435| return lenarray - low; 196| 476|} roaring.c:rle16_count_less: 157| 476| uint16_t key) { 158| 476| if (lenarray == 0) return 0; ------------------ | Branch (158:9): [True: 1, False: 475] ------------------ 159| 475| int32_t low = 0; 160| 475| int32_t high = lenarray - 1; 161| 3.01k| while (low <= high) { ------------------ | Branch (161:12): [True: 2.80k, False: 209] ------------------ 162| 2.80k| int32_t middleIndex = (low + high) >> 1; 163| 2.80k| uint16_t min_value = array[middleIndex].value; 164| 2.80k| uint16_t max_value = 165| 2.80k| array[middleIndex].value + array[middleIndex].length; 166| 2.80k| if (max_value + UINT32_C(1) < key) { // uint32 arithmetic ------------------ | Branch (166:13): [True: 1.30k, False: 1.49k] ------------------ 167| 1.30k| low = middleIndex + 1; 168| 1.49k| } else if (key < min_value) { ------------------ | Branch (168:20): [True: 1.23k, False: 266] ------------------ 169| 1.23k| high = middleIndex - 1; 170| 1.23k| } else { 171| 266| return middleIndex; 172| 266| } 173| 2.80k| } 174| 209| return low; 175| 475|} roaring.c:run_container_add_range_nruns: 610| 476| int32_t nruns_greater) { 611| 476| int32_t nruns_common = run->n_runs - nruns_less - nruns_greater; 612| 476| if (nruns_common == 0) { ------------------ | Branch (612:9): [True: 139, False: 337] ------------------ 613| 139| makeRoomAtIndex(run, (uint16_t)nruns_less); 614| 139| run->runs[nruns_less].value = (uint16_t)min; 615| 139| run->runs[nruns_less].length = (uint16_t)(max - min); 616| 337| } else { 617| 337| uint32_t common_min = run->runs[nruns_less].value; 618| 337| uint32_t common_max = run->runs[nruns_less + nruns_common - 1].value + 619| 337| run->runs[nruns_less + nruns_common - 1].length; 620| 337| uint32_t result_min = (common_min < min) ? common_min : min; ------------------ | Branch (620:31): [True: 84, False: 253] ------------------ 621| 337| uint32_t result_max = (common_max > max) ? common_max : max; ------------------ | Branch (621:31): [True: 35, False: 302] ------------------ 622| | 623| 337| run->runs[nruns_less].value = (uint16_t)result_min; 624| 337| run->runs[nruns_less].length = (uint16_t)(result_max - result_min); 625| | 626| 337| memmove(&(run->runs[nruns_less + 1]), 627| 337| &(run->runs[run->n_runs - nruns_greater]), 628| 337| nruns_greater * sizeof(rle16_t)); 629| 337| run->n_runs = nruns_less + 1 + nruns_greater; 630| 337| } 631| 476|} roaring.c:makeRoomAtIndex: 208| 821|static inline void makeRoomAtIndex(run_container_t *run, uint16_t index) { 209| | /* This function calls realloc + memmove sequentially to move by one index. 210| | * Potentially copying twice the array. 211| | */ 212| 821| if (run->n_runs + 1 > run->capacity) ------------------ | Branch (212:9): [True: 646, False: 175] ------------------ 213| 646| run_container_grow(run, run->n_runs + 1, true); 214| 821| memmove(run->runs + 1 + index, run->runs + index, 215| 821| (run->n_runs - index) * sizeof(rle16_t)); 216| 821| run->n_runs++; 217| 821|} roaring.c:run_container_remove_range: 667| 1.83k| uint32_t min, uint32_t max) { 668| 1.83k| int32_t first = rle16_find_run(run->runs, run->n_runs, (uint16_t)min); 669| 1.83k| int32_t last = rle16_find_run(run->runs, run->n_runs, (uint16_t)max); 670| | 671| 1.83k| if (first >= 0 && min > run->runs[first].value && ------------------ | Branch (671:9): [True: 653, False: 1.18k] | Branch (671:23): [True: 309, False: 344] ------------------ 672| 309| max < ((uint32_t)run->runs[first].value + ------------------ | Branch (672:9): [True: 49, False: 260] ------------------ 673| 309| (uint32_t)run->runs[first].length)) { 674| | // split this run into two adjacent runs 675| | 676| | // right subinterval 677| 49| makeRoomAtIndex(run, (uint16_t)(first + 1)); 678| 49| run->runs[first + 1].value = (uint16_t)(max + 1); 679| 49| run->runs[first + 1].length = 680| 49| (uint16_t)((run->runs[first].value + run->runs[first].length) - 681| 49| (max + 1)); 682| | 683| | // left subinterval 684| 49| run->runs[first].length = 685| 49| (uint16_t)((min - 1) - run->runs[first].value); 686| | 687| 49| return; 688| 49| } 689| | 690| | // update left-most partial run 691| 1.78k| if (first >= 0) { ------------------ | Branch (691:9): [True: 604, False: 1.18k] ------------------ 692| 604| if (min > run->runs[first].value) { ------------------ | Branch (692:13): [True: 260, False: 344] ------------------ 693| 260| run->runs[first].length = 694| 260| (uint16_t)((min - 1) - run->runs[first].value); 695| 260| first++; 696| 260| } 697| 1.18k| } else { 698| 1.18k| first = -first - 1; 699| 1.18k| } 700| | 701| | // update right-most run 702| 1.78k| if (last >= 0) { ------------------ | Branch (702:9): [True: 577, False: 1.20k] ------------------ 703| 577| uint16_t run_max = run->runs[last].value + run->runs[last].length; 704| 577| if (run_max > max) { ------------------ | Branch (704:13): [True: 335, False: 242] ------------------ 705| 335| run->runs[last].value = (uint16_t)(max + 1); 706| 335| run->runs[last].length = (uint16_t)(run_max - (max + 1)); 707| 335| last--; 708| 335| } 709| 1.20k| } else { 710| 1.20k| last = (-last - 1) - 1; 711| 1.20k| } 712| | 713| | // remove intermediate runs 714| 1.78k| if (first <= last) { ------------------ | Branch (714:9): [True: 164, False: 1.62k] ------------------ 715| 164| run_container_shift_tail(run, run->n_runs - (last + 1), 716| 164| -(last - first + 1)); 717| 164| } 718| 1.78k|} roaring.c:rle16_find_run: 133| 3.66k| uint16_t ikey) { 134| 3.66k| int32_t low = 0; 135| 3.66k| int32_t high = lenarray - 1; 136| 15.7k| while (low <= high) { ------------------ | Branch (136:12): [True: 13.3k, False: 2.38k] ------------------ 137| 13.3k| int32_t middleIndex = (low + high) >> 1; 138| 13.3k| uint16_t min = array[middleIndex].value; 139| 13.3k| uint16_t max = array[middleIndex].value + array[middleIndex].length; 140| 13.3k| if (ikey > max) { ------------------ | Branch (140:13): [True: 1.50k, False: 11.8k] ------------------ 141| 1.50k| low = middleIndex + 1; 142| 11.8k| } else if (ikey < min) { ------------------ | Branch (142:20): [True: 10.5k, False: 1.27k] ------------------ 143| 10.5k| high = middleIndex - 1; 144| 10.5k| } else { 145| 1.27k| return middleIndex; 146| 1.27k| } 147| 13.3k| } 148| 2.38k| return -(low + 1); 149| 3.66k|} roaring.c:run_container_shift_tail: 650| 164| int32_t distance) { 651| 164| if (distance > 0) { ------------------ | Branch (651:9): [True: 0, False: 164] ------------------ 652| 0| if (run->capacity < count + distance) { ------------------ | Branch (652:13): [True: 0, False: 0] ------------------ 653| | run_container_grow(run, count + distance, true); 654| 0| } 655| 0| } 656| 164| int32_t srcpos = run->n_runs - count; 657| 164| int32_t dstpos = srcpos + distance; 658| 164| memmove(&(run->runs[dstpos]), &(run->runs[srcpos]), 659| 164| sizeof(rle16_t) * count); 660| 164| run->n_runs += distance; 661| 164|} roaring.c:run_container_remove: 223| 3.14k|static inline bool run_container_remove(run_container_t *run, uint16_t pos) { 224| 3.14k| int32_t index = interleavedBinarySearch(run->runs, run->n_runs, pos); 225| 3.14k| if (index >= 0) { ------------------ | Branch (225:9): [True: 1.24k, False: 1.90k] ------------------ 226| 1.24k| int32_t le = run->runs[index].length; 227| 1.24k| if (le == 0) { ------------------ | Branch (227:13): [True: 239, False: 1.00k] ------------------ 228| 239| recoverRoomAtIndex(run, (uint16_t)index); 229| 1.00k| } else { 230| 1.00k| run->runs[index].value++; 231| 1.00k| run->runs[index].length--; 232| 1.00k| } 233| 1.24k| return true; 234| 1.24k| } 235| 1.90k| index = -index - 2; // points to preceding value, possibly -1 236| 1.90k| if (index >= 0) { // possible match ------------------ | Branch (236:9): [True: 850, False: 1.05k] ------------------ 237| 850| int32_t offset = pos - run->runs[index].value; 238| 850| int32_t le = run->runs[index].length; 239| 850| if (offset < le) { ------------------ | Branch (239:13): [True: 633, False: 217] ------------------ 240| | // need to break in two 241| 633| run->runs[index].length = (uint16_t)(offset - 1); 242| | // need to insert 243| 633| uint16_t newvalue = pos + 1; 244| 633| int32_t newlength = le - offset - 1; 245| 633| makeRoomAtIndex(run, (uint16_t)(index + 1)); 246| 633| run->runs[index + 1].value = newvalue; 247| 633| run->runs[index + 1].length = (uint16_t)newlength; 248| 633| return true; 249| | 250| 633| } else if (offset == le) { ------------------ | Branch (250:20): [True: 28, False: 189] ------------------ 251| 28| run->runs[index].length--; 252| 28| return true; 253| 28| } 254| 850| } 255| | // no match 256| 1.24k| return false; 257| 1.90k|} roaring.c:recoverRoomAtIndex: 102| 239|static inline void recoverRoomAtIndex(run_container_t *run, uint16_t index) { 103| 239| memmove(run->runs + index, run->runs + (1 + index), 104| 239| (run->n_runs - index - 1) * sizeof(rle16_t)); 105| 239| run->n_runs--; 106| 239|} roaring.c:run_container_nonzero_cardinality: 309| 61.6k| const run_container_t *run) { 310| 61.6k| return run->n_runs > 0; // runs never empty 311| 61.6k|} roaring.c:run_container_is_full: 394| 6.00k|static inline bool run_container_is_full(const run_container_t *run) { 395| 6.00k| rle16_t vl = run->runs[0]; 396| 6.00k| return (run->n_runs == 1) && (vl.value == 0) && (vl.length == 0xFFFF); ------------------ | Branch (396:12): [True: 1.15k, False: 4.84k] | Branch (396:34): [True: 894, False: 260] | Branch (396:53): [True: 772, False: 122] ------------------ 397| 6.00k|} roaring.c:run_container_equals: 502| 133| const run_container_t *container2) { 503| 133| if (container1->n_runs != container2->n_runs) { ------------------ | Branch (503:9): [True: 105, False: 28] ------------------ 504| 105| return false; 505| 105| } 506| 28| return memequals(container1->runs, container2->runs, 507| 28| container1->n_runs * sizeof(rle16_t)); 508| 133|} roaring.c:run_container_contains_range: 278| 191| uint32_t pos_end) { 279| 191| uint32_t count = 0; 280| 191| int32_t index = 281| 191| interleavedBinarySearch(run->runs, run->n_runs, (uint16_t)pos_start); 282| 191| if (index < 0) { ------------------ | Branch (282:9): [True: 124, False: 67] ------------------ 283| 124| index = -index - 2; 284| 124| if ((index == -1) || ------------------ | Branch (284:13): [True: 17, False: 107] ------------------ 285| 107| ((pos_start - run->runs[index].value) > run->runs[index].length)) { ------------------ | Branch (285:13): [True: 24, False: 83] ------------------ 286| 41| return false; 287| 41| } 288| 124| } 289| 1.81k| for (int32_t i = index; i < run->n_runs; ++i) { ------------------ | Branch (289:29): [True: 1.75k, False: 58] ------------------ 290| 1.75k| const uint32_t stop = run->runs[i].value + run->runs[i].length; 291| 1.75k| if (run->runs[i].value >= pos_end) break; ------------------ | Branch (291:13): [True: 10, False: 1.74k] ------------------ 292| 1.74k| if (stop >= pos_end) { ------------------ | Branch (292:13): [True: 82, False: 1.66k] ------------------ 293| 82| count += (((pos_end - run->runs[i].value) > 0) ------------------ | Branch (293:23): [True: 82, False: 0] ------------------ 294| 82| ? (pos_end - run->runs[i].value) 295| 82| : 0); 296| 82| break; 297| 82| } 298| 1.66k| const uint32_t min = (stop - pos_start) > 0 ? (stop - pos_start) : 0; ------------------ | Branch (298:30): [True: 1.65k, False: 9] ------------------ 299| 1.66k| count += (min < run->runs[i].length) ? min : run->runs[i].length; ------------------ | Branch (299:18): [True: 52, False: 1.61k] ------------------ 300| 1.66k| } 301| 150| return count >= (pos_end - pos_start - 1); 302| 191|} roaring_array.c:run_container_size_in_bytes: 493| 2.54k| const run_container_t *container) { 494| 2.54k| return run_container_serialized_size_in_bytes(container->n_runs); 495| 2.54k|} roaring_array.c:run_container_serialized_size_in_bytes: 456| 2.54k|static inline int32_t run_container_serialized_size_in_bytes(int32_t num_runs) { 457| 2.54k| return sizeof(uint16_t) + 458| 2.54k| sizeof(rle16_t) * num_runs; // each run requires 2 2-byte entries. 459| 2.54k|} convert.c:run_container_serialized_size_in_bytes: 456| 142k|static inline int32_t run_container_serialized_size_in_bytes(int32_t num_runs) { 457| 142k| return sizeof(uint16_t) + 458| 142k| sizeof(rle16_t) * num_runs; // each run requires 2 2-byte entries. 459| 142k|} mixed_intersection.c:run_container_is_full: 394| 16.8k|static inline bool run_container_is_full(const run_container_t *run) { 395| 16.8k| rle16_t vl = run->runs[0]; 396| 16.8k| return (run->n_runs == 1) && (vl.value == 0) && (vl.length == 0xFFFF); ------------------ | Branch (396:12): [True: 1.47k, False: 15.3k] | Branch (396:34): [True: 481, False: 997] | Branch (396:53): [True: 144, False: 337] ------------------ 397| 16.8k|} mixed_union.c:run_container_is_full: 394| 2.88k|static inline bool run_container_is_full(const run_container_t *run) { 395| 2.88k| rle16_t vl = run->runs[0]; 396| 2.88k| return (run->n_runs == 1) && (vl.value == 0) && (vl.length == 0xFFFF); ------------------ | Branch (396:12): [True: 355, False: 2.53k] | Branch (396:34): [True: 193, False: 162] | Branch (396:53): [True: 93, False: 100] ------------------ 397| 2.88k|} mixed_union.c:run_container_append_first: 351| 2.08k| rle16_t vl) { 352| 2.08k| run->runs[run->n_runs] = vl; 353| 2.08k| run->n_runs++; 354| 2.08k| return vl; 355| 2.08k|} mixed_union.c:run_container_append_value_first: 385| 705| uint16_t val) { 386| 705| rle16_t newrle = CROARING_MAKE_RLE16(val, 0); ------------------ | | 60| 705| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 387| 705| run->runs[run->n_runs] = newrle; 388| 705| run->n_runs++; 389| 705| return newrle; 390| 705|} mixed_union.c:run_container_append: 332| 170k| rle16_t *previousrl) { 333| 170k| const uint32_t previousend = previousrl->value + previousrl->length; 334| 170k| if (vl.value > previousend + 1) { // we add a new one ------------------ | Branch (334:9): [True: 157k, False: 13.0k] ------------------ 335| 157k| run->runs[run->n_runs] = vl; 336| 157k| run->n_runs++; 337| 157k| *previousrl = vl; 338| 157k| } else { 339| 13.0k| uint32_t newend = vl.value + vl.length + UINT32_C(1); 340| 13.0k| if (newend > previousend) { // we merge ------------------ | Branch (340:13): [True: 13.0k, False: 0] ------------------ 341| 13.0k| previousrl->length = (uint16_t)(newend - 1 - previousrl->value); 342| 13.0k| run->runs[run->n_runs - 1] = *previousrl; 343| 13.0k| } 344| 13.0k| } 345| 170k|} mixed_union.c:run_container_append_value: 368| 502k| rle16_t *previousrl) { 369| 502k| const uint32_t previousend = previousrl->value + previousrl->length; 370| 502k| if (val > previousend + 1) { // we add a new one ------------------ | Branch (370:9): [True: 55.3k, False: 447k] ------------------ 371| 55.3k| *previousrl = CROARING_MAKE_RLE16(val, 0); ------------------ | | 60| 55.3k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 372| 55.3k| run->runs[run->n_runs] = *previousrl; 373| 55.3k| run->n_runs++; 374| 447k| } else if (val == previousend + 1) { // we merge ------------------ | Branch (374:16): [True: 144k, False: 302k] ------------------ 375| 144k| previousrl->length++; 376| 144k| run->runs[run->n_runs - 1] = *previousrl; 377| 144k| } 378| 502k|} run.c:run_container_is_full: 394| 125k|static inline bool run_container_is_full(const run_container_t *run) { 395| 125k| rle16_t vl = run->runs[0]; 396| 125k| return (run->n_runs == 1) && (vl.value == 0) && (vl.length == 0xFFFF); ------------------ | Branch (396:12): [True: 110k, False: 14.5k] | Branch (396:34): [True: 108k, False: 1.63k] | Branch (396:53): [True: 106k, False: 2.13k] ------------------ 397| 125k|} run.c:run_container_serialized_size_in_bytes: 456| 49|static inline int32_t run_container_serialized_size_in_bytes(int32_t num_runs) { 457| 49| return sizeof(uint16_t) + 458| 49| sizeof(rle16_t) * num_runs; // each run requires 2 2-byte entries. 459| 49|} run.c:recoverRoomAtIndex: 102| 6.59k|static inline void recoverRoomAtIndex(run_container_t *run, uint16_t index) { 103| 6.59k| memmove(run->runs + index, run->runs + (1 + index), 104| 6.59k| (run->n_runs - index - 1) * sizeof(rle16_t)); 105| 6.59k| run->n_runs--; 106| 6.59k|} run.c:makeRoomAtIndex: 208| 45.7k|static inline void makeRoomAtIndex(run_container_t *run, uint16_t index) { 209| | /* This function calls realloc + memmove sequentially to move by one index. 210| | * Potentially copying twice the array. 211| | */ 212| 45.7k| if (run->n_runs + 1 > run->capacity) ------------------ | Branch (212:9): [True: 2.59k, False: 43.1k] ------------------ 213| 2.59k| run_container_grow(run, run->n_runs + 1, true); 214| 45.7k| memmove(run->runs + 1 + index, run->runs + index, 215| 45.7k| (run->n_runs - index) * sizeof(rle16_t)); 216| 45.7k| run->n_runs++; 217| 45.7k|} run.c:run_container_append_first: 351| 2.31k| rle16_t vl) { 352| 2.31k| run->runs[run->n_runs] = vl; 353| 2.31k| run->n_runs++; 354| 2.31k| return vl; 355| 2.31k|} run.c:run_container_append: 332| 303k| rle16_t *previousrl) { 333| 303k| const uint32_t previousend = previousrl->value + previousrl->length; 334| 303k| if (vl.value > previousend + 1) { // we add a new one ------------------ | Branch (334:9): [True: 153k, False: 149k] ------------------ 335| 153k| run->runs[run->n_runs] = vl; 336| 153k| run->n_runs++; 337| 153k| *previousrl = vl; 338| 153k| } else { 339| 149k| uint32_t newend = vl.value + vl.length + UINT32_C(1); 340| 149k| if (newend > previousend) { // we merge ------------------ | Branch (340:13): [True: 112k, False: 37.6k] ------------------ 341| 112k| previousrl->length = (uint16_t)(newend - 1 - previousrl->value); 342| 112k| run->runs[run->n_runs - 1] = *previousrl; 343| 112k| } 344| 149k| } 345| 303k|} run.c:run_container_empty: 314| 14|static inline bool run_container_empty(const run_container_t *run) { 315| 14| return run->n_runs == 0; // runs never empty 316| 14|} run.c:run_container_size_in_bytes: 493| 49| const run_container_t *container) { 494| 49| return run_container_serialized_size_in_bytes(container->n_runs); 495| 49|} roaring_trailing_zeroes: 242| 68.5M|inline int roaring_trailing_zeroes(unsigned long long input_num) { 243| 68.5M| return __builtin_ctzll(input_num); 244| 68.5M|} roaring_leading_zeroes: 246| 4.23k|inline int roaring_leading_zeroes(unsigned long long input_num) { 247| 4.23k| return __builtin_clzll(input_num); 248| 4.23k|} roaring.c:roaring_hamming: 287| 153k|static inline int roaring_hamming(uint64_t x) { 288| |#if defined(_WIN64) && defined(CROARING_REGULAR_VISUAL_STUDIO) && \ 289| | CROARING_REGULAR_VISUAL_STUDIO 290| |#ifdef CROARING_USENEON 291| | return vaddv_u8(vcnt_u8(vcreate_u8(input_num))); 292| |#elif defined(_M_ARM64) 293| | return roaring_hamming_backup(x); 294| | // (int) _CountOneBits64(x); is unavailable 295| |#else // _M_ARM64 296| | return (int)__popcnt64(x); 297| |#endif // _M_ARM64 298| |#elif defined(_WIN32) && defined(CROARING_REGULAR_VISUAL_STUDIO) && \ 299| | CROARING_REGULAR_VISUAL_STUDIO 300| |#ifdef _M_ARM 301| | return roaring_hamming_backup(x); 302| | // _CountOneBits is unavailable 303| |#else // _M_ARM 304| | return (int)__popcnt((unsigned int)x) + 305| | (int)__popcnt((unsigned int)(x >> 32)); 306| |#endif // _M_ARM 307| |#else 308| 153k| return __builtin_popcountll(x); 309| 153k|#endif 310| 153k|} mixed_intersection.c:roaring_hamming: 287| 207k|static inline int roaring_hamming(uint64_t x) { 288| |#if defined(_WIN64) && defined(CROARING_REGULAR_VISUAL_STUDIO) && \ 289| | CROARING_REGULAR_VISUAL_STUDIO 290| |#ifdef CROARING_USENEON 291| | return vaddv_u8(vcnt_u8(vcreate_u8(input_num))); 292| |#elif defined(_M_ARM64) 293| | return roaring_hamming_backup(x); 294| | // (int) _CountOneBits64(x); is unavailable 295| |#else // _M_ARM64 296| | return (int)__popcnt64(x); 297| |#endif // _M_ARM64 298| |#elif defined(_WIN32) && defined(CROARING_REGULAR_VISUAL_STUDIO) && \ 299| | CROARING_REGULAR_VISUAL_STUDIO 300| |#ifdef _M_ARM 301| | return roaring_hamming_backup(x); 302| | // _CountOneBits is unavailable 303| |#else // _M_ARM 304| | return (int)__popcnt((unsigned int)x) + 305| | (int)__popcnt((unsigned int)(x >> 32)); 306| |#endif // _M_ARM 307| |#else 308| 207k| return __builtin_popcountll(x); 309| 207k|#endif 310| 207k|} roaring_bitmap_init_cleared: 72| 31.2k|inline void roaring_bitmap_init_cleared(roaring_bitmap_t *r) { 73| 31.2k| roaring_bitmap_init_with_capacity(r, 0); 74| 31.2k|} roaring_bitmap_add_range: 434| 6.24k| uint64_t max) { 435| 6.24k| if (max <= min || min > (uint64_t)UINT32_MAX + 1) { ------------------ | Branch (435:9): [True: 3.53k, False: 2.71k] | Branch (435:23): [True: 0, False: 2.71k] ------------------ 436| 3.53k| return; 437| 3.53k| } 438| 2.71k| roaring_bitmap_add_range_closed(r, (uint32_t)min, (uint32_t)(max - 1)); 439| 2.71k|} roaring_bitmap_remove_range: 456| 6.24k| uint64_t max) { 457| 6.24k| if (max <= min || min > (uint64_t)UINT32_MAX + 1) { ------------------ | Branch (457:9): [True: 5.35k, False: 896] | Branch (457:23): [True: 0, False: 896] ------------------ 458| 5.35k| return; 459| 5.35k| } 460| 896| roaring_bitmap_remove_range_closed(r, (uint32_t)min, (uint32_t)(max - 1)); 461| 896|} roaring_bitmap_contains: 478| 6.29k|inline bool roaring_bitmap_contains(const roaring_bitmap_t *r, uint32_t val) { 479| | // For performance reasons, this function is inline and uses internal 480| | // functions directly. 481| 6.29k|#ifdef __cplusplus 482| 6.29k| using namespace ::roaring::internal; 483| 6.29k|#endif 484| 6.29k| const uint16_t hb = val >> 16; 485| | /* 486| | * the next function call involves a binary search and lots of branching. 487| | */ 488| 6.29k| int32_t i = ra_get_index(&r->high_low_container, hb); 489| 6.29k| if (i < 0) return false; ------------------ | Branch (489:9): [True: 632, False: 5.66k] ------------------ 490| | 491| 5.66k| uint8_t typecode; 492| | // next call ought to be cheap 493| 5.66k| container_t *container = ra_get_container_at_index(&r->high_low_container, 494| 5.66k| (uint16_t)i, &typecode); 495| | // rest might be a tad expensive, possibly involving another round of binary 496| | // search 497| 5.66k| return container_contains(container, val & 0xFFFF, typecode); 498| 6.29k|} roaring_bitmap_set_copy_on_write: 117| 57.7k|inline void roaring_bitmap_set_copy_on_write(roaring_bitmap_t *r, bool cow) { 118| 57.7k| if (cow) { ------------------ | Branch (118:9): [True: 0, False: 57.7k] ------------------ 119| 0| r->high_low_container.flags |= ROARING_FLAG_COW; ------------------ | | 46| 0|#define ROARING_FLAG_COW UINT8_C(0x1) ------------------ 120| 57.7k| } else { 121| 57.7k| if (roaring_bitmap_get_copy_on_write(r)) { ------------------ | Branch (121:13): [True: 0, False: 57.7k] ------------------ 122| 0| roaring_unshare_all(r); 123| 0| } 124| | r->high_low_container.flags &= ~ROARING_FLAG_COW; ------------------ | | 46| 57.7k|#define ROARING_FLAG_COW UINT8_C(0x1) ------------------ 125| 57.7k| } 126| 57.7k|} roaring_bitmap_get_copy_on_write: 114| 70.2k|inline bool roaring_bitmap_get_copy_on_write(const roaring_bitmap_t *r) { 115| | return r->high_low_container.flags & ROARING_FLAG_COW; ------------------ | | 46| 70.2k|#define ROARING_FLAG_COW UINT8_C(0x1) ------------------ 116| 70.2k|} ra_get_index: 93| 87.4k|inline int32_t ra_get_index(const roaring_array_t *ra, uint16_t x) { 94| 87.4k| if ((ra->size == 0) || ra->keys[ra->size - 1] == x) return ra->size - 1; ------------------ | Branch (94:9): [True: 12.6k, False: 74.7k] | Branch (94:28): [True: 32.6k, False: 42.1k] ------------------ 95| 42.1k| return binarySearch(ra->keys, (int32_t)ra->size, x); 96| 87.4k|} ra_get_container_at_index: 102| 435k| uint16_t i, uint8_t *typecode) { 103| 435k| *typecode = ra->typecodes[i]; 104| 435k| return ra->containers[i]; 105| 435k|} ra_get_key_at_index: 110| 297k|inline uint16_t ra_get_key_at_index(const roaring_array_t *ra, uint16_t i) { 111| 297k| return ra->keys[i]; 112| 297k|} ra_set_container_at_index: 175| 108k| container_t *c, uint8_t typecode) { 176| | assert(i < ra->size); 177| 108k| ra->containers[i] = c; 178| 108k| ra->typecodes[i] = typecode; 179| 108k|} ra_get_size: 191| 13.1k|inline int32_t ra_get_size(const roaring_array_t *ra) { return ra->size; } ra_replace_key_and_container_at_index: 205| 145k| uint8_t typecode) { 206| 145k| assert(i < ra->size); 207| | 208| 145k| ra->keys[i] = key; 209| 145k| ra->containers[i] = c; 210| 145k| ra->typecodes[i] = typecode; 211| 145k|} roaring.c:ra_advance_until: 194| 469| int32_t pos) { 195| 469| return advanceUntil(ra->keys, pos, ra->size, x); 196| 469|} roaring.c:ra_unshare_container_at_index: 268| 63.0k| uint16_t i) { 269| | assert(i < ra->size); 270| 63.0k| ra->containers[i] = 271| 63.0k| get_writable_copy_if_shared(ra->containers[i], &ra->typecodes[i]); 272| 63.0k|} intersect_vector16: 386| 1.76k| size_t s_b, uint16_t *C) { 387| 1.76k| size_t count = 0; 388| 1.76k| size_t i_a = 0, i_b = 0; 389| 1.76k| const int vectorlength = sizeof(__m128i) / sizeof(uint16_t); 390| 1.76k| const size_t st_a = (s_a / vectorlength) * vectorlength; 391| 1.76k| const size_t st_b = (s_b / vectorlength) * vectorlength; 392| 1.76k| __m128i v_a, v_b; 393| 1.76k| if ((i_a < st_a) && (i_b < st_b)) { ------------------ | Branch (393:9): [True: 1.63k, False: 135] | Branch (393:25): [True: 1.48k, False: 149] ------------------ 394| 1.48k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 395| 1.48k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 396| 4.01k| while ((A[i_a] == 0) || (B[i_b] == 0)) { ------------------ | Branch (396:16): [True: 2.52k, False: 1.49k] | Branch (396:33): [True: 182, False: 1.31k] ------------------ 397| 2.70k| const __m128i res_v = _mm_cmpestrm( 398| 2.70k| v_b, vectorlength, v_a, vectorlength, 399| 2.70k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 400| 2.70k| const int r = _mm_extract_epi32(res_v, 0); 401| 2.70k| __m128i sm16 = _mm_loadu_si128((const __m128i *)shuffle_mask16 + r); 402| 2.70k| __m128i p = _mm_shuffle_epi8(v_a, sm16); 403| 2.70k| _mm_storeu_si128((__m128i *)&C[count], p); // can overflow 404| 2.70k| count += _mm_popcnt_u32(r); 405| 2.70k| const uint16_t a_max = A[i_a + vectorlength - 1]; 406| 2.70k| const uint16_t b_max = B[i_b + vectorlength - 1]; 407| 2.70k| if (a_max <= b_max) { ------------------ | Branch (407:17): [True: 1.51k, False: 1.19k] ------------------ 408| 1.51k| i_a += vectorlength; 409| 1.51k| if (i_a == st_a) break; ------------------ | Branch (409:21): [True: 130, False: 1.38k] ------------------ 410| 1.38k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 411| 1.38k| } 412| 2.57k| if (b_max <= a_max) { ------------------ | Branch (412:17): [True: 1.25k, False: 1.31k] ------------------ 413| 1.25k| i_b += vectorlength; 414| 1.25k| if (i_b == st_b) break; ------------------ | Branch (414:21): [True: 35, False: 1.22k] ------------------ 415| 1.22k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 416| 1.22k| } 417| 2.57k| } 418| 1.48k| if ((i_a < st_a) && (i_b < st_b)) ------------------ | Branch (418:13): [True: 1.35k, False: 130] | Branch (418:29): [True: 1.31k, False: 35] ------------------ 419| 41.2k| while (true) { ------------------ | Branch (419:20): [True: 41.2k, Folded] ------------------ 420| 41.2k| const __m128i res_v = _mm_cmpistrm( 421| 41.2k| v_b, v_a, 422| 41.2k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 423| 41.2k| const int r = _mm_extract_epi32(res_v, 0); 424| 41.2k| __m128i sm16 = 425| 41.2k| _mm_loadu_si128((const __m128i *)shuffle_mask16 + r); 426| 41.2k| __m128i p = _mm_shuffle_epi8(v_a, sm16); 427| 41.2k| _mm_storeu_si128((__m128i *)&C[count], p); // can overflow 428| 41.2k| count += _mm_popcnt_u32(r); 429| 41.2k| const uint16_t a_max = A[i_a + vectorlength - 1]; 430| 41.2k| const uint16_t b_max = B[i_b + vectorlength - 1]; 431| 41.2k| if (a_max <= b_max) { ------------------ | Branch (431:21): [True: 18.5k, False: 22.7k] ------------------ 432| 18.5k| i_a += vectorlength; 433| 18.5k| if (i_a == st_a) break; ------------------ | Branch (433:25): [True: 755, False: 17.7k] ------------------ 434| 17.7k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 435| 17.7k| } 436| 40.5k| if (b_max <= a_max) { ------------------ | Branch (436:21): [True: 24.2k, False: 16.3k] ------------------ 437| 24.2k| i_b += vectorlength; 438| 24.2k| if (i_b == st_b) break; ------------------ | Branch (438:25): [True: 562, False: 23.6k] ------------------ 439| 23.6k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 440| 23.6k| } 441| 40.5k| } 442| 1.48k| } 443| | // intersect the tail using scalar intersection 444| 28.7k| while (i_a < s_a && i_b < s_b) { ------------------ | Branch (444:12): [True: 27.3k, False: 1.38k] | Branch (444:25): [True: 26.9k, False: 378] ------------------ 445| 26.9k| uint16_t a = A[i_a]; 446| 26.9k| uint16_t b = B[i_b]; 447| 26.9k| if (a < b) { ------------------ | Branch (447:13): [True: 11.4k, False: 15.5k] ------------------ 448| 11.4k| i_a++; 449| 15.5k| } else if (b < a) { ------------------ | Branch (449:20): [True: 9.95k, False: 5.57k] ------------------ 450| 9.95k| i_b++; 451| 9.95k| } else { 452| 5.57k| C[count] = a; //==b; 453| 5.57k| count++; 454| 5.57k| i_a++; 455| 5.57k| i_b++; 456| 5.57k| } 457| 26.9k| } 458| 1.76k| return (int32_t)count; 459| 1.76k|} array_container_to_uint32_array_vector16: 464| 5.62k| uint32_t base) { 465| 5.62k| int outpos = 0; 466| 5.62k| uint32_t *out = (uint32_t *)vout; 467| 5.62k| size_t i = 0; 468| 183k| for (; i + sizeof(__m128i) / sizeof(uint16_t) <= cardinality; ------------------ | Branch (468:12): [True: 177k, False: 5.62k] ------------------ 469| 177k| i += sizeof(__m128i) / sizeof(uint16_t)) { 470| 177k| __m128i vinput = _mm_loadu_si128((const __m128i *)(array + i)); 471| 177k| __m256i voutput = _mm256_add_epi32(_mm256_cvtepu16_epi32(vinput), 472| 177k| _mm256_set1_epi32(base)); 473| 177k| _mm256_storeu_si256((__m256i *)(out + outpos), voutput); 474| 177k| outpos += sizeof(__m256i) / sizeof(uint32_t); 475| 177k| } 476| 19.7k| for (; i < cardinality; ++i) { ------------------ | Branch (476:12): [True: 14.1k, False: 5.62k] ------------------ 477| 14.1k| const uint32_t val = base + array[i]; 478| 14.1k| memcpy(out + outpos, &val, 479| 14.1k| sizeof(uint32_t)); // should be compiled as a MOV on x64 480| 14.1k| outpos++; 481| 14.1k| } 482| 5.62k| return outpos; 483| 5.62k|} intersect_vector16_inplace: 486| 4.36k| size_t s_b) { 487| 4.36k| size_t count = 0; 488| 4.36k| size_t i_a = 0, i_b = 0; 489| 4.36k| const int vectorlength = sizeof(__m128i) / sizeof(uint16_t); 490| 4.36k| const size_t st_a = (s_a / vectorlength) * vectorlength; 491| 4.36k| const size_t st_b = (s_b / vectorlength) * vectorlength; 492| 4.36k| __m128i v_a, v_b; 493| 4.36k| if ((i_a < st_a) && (i_b < st_b)) { ------------------ | Branch (493:9): [True: 2.26k, False: 2.10k] | Branch (493:25): [True: 2.07k, False: 185] ------------------ 494| 2.07k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 495| 2.07k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 496| 2.07k| __m128i tmp[2] = {_mm_setzero_si128()}; 497| 2.07k| size_t tmp_count = 0; 498| 4.08k| while ((A[i_a] == 0) || (B[i_b] == 0)) { ------------------ | Branch (498:16): [True: 1.83k, False: 2.25k] | Branch (498:33): [True: 286, False: 1.96k] ------------------ 499| 2.11k| const __m128i res_v = _mm_cmpestrm( 500| 2.11k| v_b, vectorlength, v_a, vectorlength, 501| 2.11k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 502| 2.11k| const int r = _mm_extract_epi32(res_v, 0); 503| 2.11k| __m128i sm16 = _mm_loadu_si128((const __m128i *)shuffle_mask16 + r); 504| 2.11k| __m128i p = _mm_shuffle_epi8(v_a, sm16); 505| 2.11k| _mm_storeu_si128((__m128i *)&((uint16_t *)tmp)[tmp_count], p); 506| 2.11k| tmp_count += _mm_popcnt_u32(r); 507| 2.11k| const uint16_t a_max = A[i_a + vectorlength - 1]; 508| 2.11k| const uint16_t b_max = B[i_b + vectorlength - 1]; 509| 2.11k| if (a_max <= b_max) { ------------------ | Branch (509:17): [True: 1.98k, False: 136] ------------------ 510| 1.98k| _mm_storeu_si128((__m128i *)&A[count], tmp[0]); 511| 1.98k| _mm_storeu_si128(tmp, _mm_setzero_si128()); 512| 1.98k| count += tmp_count; 513| 1.98k| tmp_count = 0; 514| 1.98k| i_a += vectorlength; 515| 1.98k| if (i_a == st_a) break; ------------------ | Branch (515:21): [True: 105, False: 1.87k] ------------------ 516| 1.87k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 517| 1.87k| } 518| 2.01k| if (b_max <= a_max) { ------------------ | Branch (518:17): [True: 327, False: 1.68k] ------------------ 519| 327| i_b += vectorlength; 520| 327| if (i_b == st_b) break; ------------------ | Branch (520:21): [True: 9, False: 318] ------------------ 521| 318| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 522| 318| } 523| 2.01k| } 524| 2.07k| if ((i_a < st_a) && (i_b < st_b)) { ------------------ | Branch (524:13): [True: 1.97k, False: 105] | Branch (524:29): [True: 1.96k, False: 9] ------------------ 525| 220k| while (true) { ------------------ | Branch (525:20): [True: 220k, Folded] ------------------ 526| 220k| const __m128i res_v = _mm_cmpistrm( 527| 220k| v_b, v_a, 528| 220k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 529| 220k| const int r = _mm_extract_epi32(res_v, 0); 530| 220k| __m128i sm16 = 531| 220k| _mm_loadu_si128((const __m128i *)shuffle_mask16 + r); 532| 220k| __m128i p = _mm_shuffle_epi8(v_a, sm16); 533| 220k| _mm_storeu_si128((__m128i *)&((uint16_t *)tmp)[tmp_count], p); 534| 220k| tmp_count += _mm_popcnt_u32(r); 535| 220k| const uint16_t a_max = A[i_a + vectorlength - 1]; 536| 220k| const uint16_t b_max = B[i_b + vectorlength - 1]; 537| 220k| if (a_max <= b_max) { ------------------ | Branch (537:21): [True: 129k, False: 91.0k] ------------------ 538| 129k| _mm_storeu_si128((__m128i *)&A[count], tmp[0]); 539| 129k| _mm_storeu_si128(tmp, _mm_setzero_si128()); 540| 129k| count += tmp_count; 541| 129k| tmp_count = 0; 542| 129k| i_a += vectorlength; 543| 129k| if (i_a == st_a) break; ------------------ | Branch (543:25): [True: 1.05k, False: 128k] ------------------ 544| 128k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 545| 128k| } 546| 219k| if (b_max <= a_max) { ------------------ | Branch (546:21): [True: 123k, False: 95.9k] ------------------ 547| 123k| i_b += vectorlength; 548| 123k| if (i_b == st_b) break; ------------------ | Branch (548:25): [True: 912, False: 122k] ------------------ 549| 122k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 550| 122k| } 551| 219k| } 552| 1.96k| } 553| | // tmp_count <= 8, so this does not affect efficiency so much 554| 4.57k| for (size_t i = 0; i < tmp_count; i++) { ------------------ | Branch (554:28): [True: 2.49k, False: 2.07k] ------------------ 555| 2.49k| A[count] = ((uint16_t *)tmp)[i]; 556| 2.49k| count++; 557| 2.49k| } 558| 2.07k| i_a += tmp_count; // We can at least jump pass $tmp_count elements in A 559| 2.07k| } 560| | // intersect the tail using scalar intersection 561| 33.2k| while (i_a < s_a && i_b < s_b) { ------------------ | Branch (561:12): [True: 29.3k, False: 3.93k] | Branch (561:25): [True: 28.8k, False: 433] ------------------ 562| 28.8k| uint16_t a = A[i_a]; 563| 28.8k| uint16_t b = B[i_b]; 564| 28.8k| if (a < b) { ------------------ | Branch (564:13): [True: 10.2k, False: 18.6k] ------------------ 565| 10.2k| i_a++; 566| 18.6k| } else if (b < a) { ------------------ | Branch (566:20): [True: 7.06k, False: 11.5k] ------------------ 567| 7.06k| i_b++; 568| 11.5k| } else { 569| 11.5k| A[count] = a; //==b; 570| 11.5k| count++; 571| 11.5k| i_a++; 572| 11.5k| i_b++; 573| 11.5k| } 574| 28.8k| } 575| 4.36k| return (int32_t)count; 576| 4.36k|} intersect_vector16_cardinality: 581| 7.06k| const uint16_t *B, size_t s_b) { 582| 7.06k| size_t count = 0; 583| 7.06k| size_t i_a = 0, i_b = 0; 584| 7.06k| const int vectorlength = sizeof(__m128i) / sizeof(uint16_t); 585| 7.06k| const size_t st_a = (s_a / vectorlength) * vectorlength; 586| 7.06k| const size_t st_b = (s_b / vectorlength) * vectorlength; 587| 7.06k| __m128i v_a, v_b; 588| 7.06k| if ((i_a < st_a) && (i_b < st_b)) { ------------------ | Branch (588:9): [True: 6.52k, False: 540] | Branch (588:25): [True: 5.92k, False: 596] ------------------ 589| 5.92k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 590| 5.92k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 591| 16.0k| while ((A[i_a] == 0) || (B[i_b] == 0)) { ------------------ | Branch (591:16): [True: 10.0k, False: 5.99k] | Branch (591:33): [True: 728, False: 5.26k] ------------------ 592| 10.8k| const __m128i res_v = _mm_cmpestrm( 593| 10.8k| v_b, vectorlength, v_a, vectorlength, 594| 10.8k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 595| 10.8k| const int r = _mm_extract_epi32(res_v, 0); 596| 10.8k| count += _mm_popcnt_u32(r); 597| 10.8k| const uint16_t a_max = A[i_a + vectorlength - 1]; 598| 10.8k| const uint16_t b_max = B[i_b + vectorlength - 1]; 599| 10.8k| if (a_max <= b_max) { ------------------ | Branch (599:17): [True: 6.04k, False: 4.76k] ------------------ 600| 6.04k| i_a += vectorlength; 601| 6.04k| if (i_a == st_a) break; ------------------ | Branch (601:21): [True: 520, False: 5.52k] ------------------ 602| 5.52k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 603| 5.52k| } 604| 10.2k| if (b_max <= a_max) { ------------------ | Branch (604:17): [True: 5.03k, False: 5.25k] ------------------ 605| 5.03k| i_b += vectorlength; 606| 5.03k| if (i_b == st_b) break; ------------------ | Branch (606:21): [True: 140, False: 4.89k] ------------------ 607| 4.89k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 608| 4.89k| } 609| 10.2k| } 610| 5.92k| if ((i_a < st_a) && (i_b < st_b)) ------------------ | Branch (610:13): [True: 5.40k, False: 520] | Branch (610:29): [True: 5.26k, False: 140] ------------------ 611| 165k| while (true) { ------------------ | Branch (611:20): [True: 165k, Folded] ------------------ 612| 165k| const __m128i res_v = _mm_cmpistrm( 613| 165k| v_b, v_a, 614| 165k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 615| 165k| const int r = _mm_extract_epi32(res_v, 0); 616| 165k| count += _mm_popcnt_u32(r); 617| 165k| const uint16_t a_max = A[i_a + vectorlength - 1]; 618| 165k| const uint16_t b_max = B[i_b + vectorlength - 1]; 619| 165k| if (a_max <= b_max) { ------------------ | Branch (619:21): [True: 74.0k, False: 91.1k] ------------------ 620| 74.0k| i_a += vectorlength; 621| 74.0k| if (i_a == st_a) break; ------------------ | Branch (621:25): [True: 3.02k, False: 71.0k] ------------------ 622| 71.0k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 623| 71.0k| } 624| 162k| if (b_max <= a_max) { ------------------ | Branch (624:21): [True: 96.8k, False: 65.2k] ------------------ 625| 96.8k| i_b += vectorlength; 626| 96.8k| if (i_b == st_b) break; ------------------ | Branch (626:25): [True: 2.24k, False: 94.6k] ------------------ 627| 94.6k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 628| 94.6k| } 629| 162k| } 630| 5.92k| } 631| | // intersect the tail using scalar intersection 632| 114k| while (i_a < s_a && i_b < s_b) { ------------------ | Branch (632:12): [True: 109k, False: 5.55k] | Branch (632:25): [True: 107k, False: 1.51k] ------------------ 633| 107k| uint16_t a = A[i_a]; 634| 107k| uint16_t b = B[i_b]; 635| 107k| if (a < b) { ------------------ | Branch (635:13): [True: 45.6k, False: 62.1k] ------------------ 636| 45.6k| i_a++; 637| 62.1k| } else if (b < a) { ------------------ | Branch (637:20): [True: 39.8k, False: 22.2k] ------------------ 638| 39.8k| i_b++; 639| 39.8k| } else { 640| 22.2k| count++; 641| 22.2k| i_a++; 642| 22.2k| i_b++; 643| 22.2k| } 644| 107k| } 645| 7.06k| return (int32_t)count; 646| 7.06k|} difference_vector16: 655| 2.30k| size_t s_b, uint16_t *C) { 656| | // we handle the degenerate case 657| 2.30k| if (s_a == 0) return 0; ------------------ | Branch (657:9): [True: 0, False: 2.30k] ------------------ 658| 2.30k| if (s_b == 0) { ------------------ | Branch (658:9): [True: 0, False: 2.30k] ------------------ 659| 0| if (A != C) memcpy(C, A, sizeof(uint16_t) * s_a); ------------------ | Branch (659:13): [True: 0, False: 0] ------------------ 660| 0| return (int32_t)s_a; 661| 0| } 662| | // handle the leading zeroes, it is messy but it allows us to use the fast 663| | // _mm_cmpistrm instrinsic safely 664| 2.30k| int32_t count = 0; 665| 2.30k| if ((A[0] == 0) || (B[0] == 0)) { ------------------ | Branch (665:9): [True: 2.11k, False: 186] | Branch (665:24): [True: 4, False: 182] ------------------ 666| 2.12k| if ((A[0] == 0) && (B[0] == 0)) { ------------------ | Branch (666:13): [True: 2.11k, False: 4] | Branch (666:28): [True: 291, False: 1.82k] ------------------ 667| 291| A++; 668| 291| s_a--; 669| 291| B++; 670| 291| s_b--; 671| 1.83k| } else if (A[0] == 0) { ------------------ | Branch (671:20): [True: 1.82k, False: 4] ------------------ 672| 1.82k| C[count++] = 0; 673| 1.82k| A++; 674| 1.82k| s_a--; 675| 1.82k| } else { 676| 4| B++; 677| 4| s_b--; 678| 4| } 679| 2.12k| } 680| | // at this point, we have two non-empty arrays, made of non-zero 681| | // increasing values. 682| 2.30k| size_t i_a = 0, i_b = 0; 683| 2.30k| const size_t vectorlength = sizeof(__m128i) / sizeof(uint16_t); 684| 2.30k| const size_t st_a = (s_a / vectorlength) * vectorlength; 685| 2.30k| const size_t st_b = (s_b / vectorlength) * vectorlength; 686| 2.30k| if ((i_a < st_a) && (i_b < st_b)) { // this is the vectorized code path ------------------ | Branch (686:9): [True: 1.89k, False: 411] | Branch (686:25): [True: 1.60k, False: 293] ------------------ 687| 1.60k| __m128i v_a, v_b; //, v_bmax; 688| | // we load a vector from A and a vector from B 689| 1.60k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 690| 1.60k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 691| | // we have a runningmask which indicates which values from A have been 692| | // spotted in B, these don't get written out. 693| 1.60k| __m128i runningmask_a_found_in_b = _mm_setzero_si128(); 694| | /**** 695| | * start of the main vectorized loop 696| | *****/ 697| 45.1k| while (true) { ------------------ | Branch (697:16): [True: 45.1k, Folded] ------------------ 698| | // afoundinb will contain a mask indicate for each entry in A 699| | // whether it is seen 700| | // in B 701| 45.1k| const __m128i a_found_in_b = _mm_cmpistrm( 702| 45.1k| v_b, v_a, 703| 45.1k| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 704| 45.1k| runningmask_a_found_in_b = 705| 45.1k| _mm_or_si128(runningmask_a_found_in_b, a_found_in_b); 706| | // we always compare the last values of A and B 707| 45.1k| const uint16_t a_max = A[i_a + vectorlength - 1]; 708| 45.1k| const uint16_t b_max = B[i_b + vectorlength - 1]; 709| 45.1k| if (a_max <= b_max) { ------------------ | Branch (709:17): [True: 20.4k, False: 24.7k] ------------------ 710| | // Ok. In this code path, we are ready to write our v_a 711| | // because there is no need to read more from B, they will 712| | // all be large values. 713| 20.4k| const int bitmask_belongs_to_difference = 714| 20.4k| _mm_extract_epi32(runningmask_a_found_in_b, 0) ^ 0xFF; 715| | /*** next few lines are probably expensive *****/ 716| 20.4k| __m128i sm16 = _mm_loadu_si128((const __m128i *)shuffle_mask16 + 717| 20.4k| bitmask_belongs_to_difference); 718| 20.4k| __m128i p = _mm_shuffle_epi8(v_a, sm16); 719| 20.4k| _mm_storeu_si128((__m128i *)&C[count], p); // can overflow 720| 20.4k| count += _mm_popcnt_u32(bitmask_belongs_to_difference); 721| | // we advance a 722| 20.4k| i_a += vectorlength; 723| 20.4k| if (i_a == st_a) // no more ------------------ | Branch (723:21): [True: 1.12k, False: 19.2k] ------------------ 724| 1.12k| break; 725| 19.2k| runningmask_a_found_in_b = _mm_setzero_si128(); 726| 19.2k| v_a = _mm_lddqu_si128((__m128i *)&A[i_a]); 727| 19.2k| } 728| 44.0k| if (b_max <= a_max) { ------------------ | Branch (728:17): [True: 27.7k, False: 16.3k] ------------------ 729| | // in this code path, the current v_b has become useless 730| 27.7k| i_b += vectorlength; 731| 27.7k| if (i_b == st_b) break; ------------------ | Branch (731:21): [True: 474, False: 27.2k] ------------------ 732| 27.2k| v_b = _mm_lddqu_si128((__m128i *)&B[i_b]); 733| 27.2k| } 734| 44.0k| } 735| | // at this point, either we have i_a == st_a, which is the end of the 736| | // vectorized processing, 737| | // or we have i_b == st_b, and we are not done processing the vector... 738| | // so we need to finish it off. 739| 1.60k| if (i_a < st_a) { // we have unfinished business... ------------------ | Branch (739:13): [True: 474, False: 1.12k] ------------------ 740| 474| uint16_t buffer[8]; // buffer to do a masked load 741| 474| memset(buffer, 0, 8 * sizeof(uint16_t)); 742| 474| memcpy(buffer, B + i_b, (s_b - i_b) * sizeof(uint16_t)); 743| 474| v_b = _mm_lddqu_si128((__m128i *)buffer); 744| 474| const __m128i a_found_in_b = _mm_cmpistrm( 745| 474| v_b, v_a, 746| 474| _SIDD_UWORD_OPS | _SIDD_CMP_EQUAL_ANY | _SIDD_BIT_MASK); 747| 474| runningmask_a_found_in_b = 748| 474| _mm_or_si128(runningmask_a_found_in_b, a_found_in_b); 749| 474| const int bitmask_belongs_to_difference = 750| 474| _mm_extract_epi32(runningmask_a_found_in_b, 0) ^ 0xFF; 751| 474| __m128i sm16 = _mm_loadu_si128((const __m128i *)shuffle_mask16 + 752| 474| bitmask_belongs_to_difference); 753| 474| __m128i p = _mm_shuffle_epi8(v_a, sm16); 754| 474| _mm_storeu_si128((__m128i *)&C[count], p); // can overflow 755| 474| count += _mm_popcnt_u32(bitmask_belongs_to_difference); 756| 474| i_a += vectorlength; 757| 474| } 758| | // at this point we should have i_a == st_a and i_b == st_b 759| 1.60k| } 760| | // do the tail using scalar code 761| 107k| while (i_a < s_a && i_b < s_b) { ------------------ | Branch (761:12): [True: 105k, False: 1.78k] | Branch (761:25): [True: 105k, False: 520] ------------------ 762| 105k| uint16_t a = A[i_a]; 763| 105k| uint16_t b = B[i_b]; 764| 105k| if (b < a) { ------------------ | Branch (764:13): [True: 72.9k, False: 32.1k] ------------------ 765| 72.9k| i_b++; 766| 72.9k| } else if (a < b) { ------------------ | Branch (766:20): [True: 26.6k, False: 5.51k] ------------------ 767| 26.6k| C[count] = a; 768| 26.6k| count++; 769| 26.6k| i_a++; 770| 26.6k| } else { //== 771| 5.51k| i_a++; 772| 5.51k| i_b++; 773| 5.51k| } 774| 105k| } 775| 2.30k| if (i_a < s_a) { ------------------ | Branch (775:9): [True: 520, False: 1.78k] ------------------ 776| 520| if (C == A) { ------------------ | Branch (776:13): [True: 0, False: 520] ------------------ 777| 0| assert((size_t)count <= i_a); 778| 0| if ((size_t)count < i_a) { ------------------ | Branch (778:17): [True: 0, False: 0] ------------------ 779| 0| memmove(C + count, A + i_a, sizeof(uint16_t) * (s_a - i_a)); 780| 0| } 781| 520| } else { 782| 26.8k| for (size_t i = 0; i < (s_a - i_a); i++) { ------------------ | Branch (782:32): [True: 26.3k, False: 520] ------------------ 783| 26.3k| C[count + i] = A[i + i_a]; 784| 26.3k| } 785| 520| } 786| 520| count += (int32_t)(s_a - i_a); 787| 520| } 788| 2.30k| return count; 789| 2.30k|} intersect_skewed_uint16: 855| 743| uint16_t *buffer) { 856| 743| size_t pos = 0, idx_l = 0, idx_s = 0; 857| | 858| 743| if (0 == size_s) { ------------------ | Branch (858:9): [True: 0, False: 743] ------------------ 859| 0| return 0; 860| 0| } 861| 743| int32_t index1 = 0, index2 = 0, index3 = 0, index4 = 0; 862| 2.10k| while ((idx_s + 4 <= size_s) && (idx_l < size_l)) { ------------------ | Branch (862:12): [True: 1.36k, False: 742] | Branch (862:37): [True: 1.36k, False: 1] ------------------ 863| 1.36k| uint16_t target1 = small[idx_s]; 864| 1.36k| uint16_t target2 = small[idx_s + 1]; 865| 1.36k| uint16_t target3 = small[idx_s + 2]; 866| 1.36k| uint16_t target4 = small[idx_s + 3]; 867| 1.36k| binarySearch4(large + idx_l, (int32_t)(size_l - idx_l), target1, 868| 1.36k| target2, target3, target4, &index1, &index2, &index3, 869| 1.36k| &index4); 870| 1.36k| if ((index1 + idx_l < size_l) && (large[idx_l + index1] == target1)) { ------------------ | Branch (870:13): [True: 1.36k, False: 1] | Branch (870:42): [True: 681, False: 680] ------------------ 871| 681| buffer[pos++] = target1; 872| 681| } 873| 1.36k| if ((index2 + idx_l < size_l) && (large[idx_l + index2] == target2)) { ------------------ | Branch (873:13): [True: 1.36k, False: 2] | Branch (873:42): [True: 841, False: 519] ------------------ 874| 841| buffer[pos++] = target2; 875| 841| } 876| 1.36k| if ((index3 + idx_l < size_l) && (large[idx_l + index3] == target3)) { ------------------ | Branch (876:13): [True: 1.35k, False: 3] | Branch (876:42): [True: 822, False: 537] ------------------ 877| 822| buffer[pos++] = target3; 878| 822| } 879| 1.36k| if ((index4 + idx_l < size_l) && (large[idx_l + index4] == target4)) { ------------------ | Branch (879:13): [True: 1.35k, False: 6] | Branch (879:42): [True: 804, False: 552] ------------------ 880| 804| buffer[pos++] = target4; 881| 804| } 882| 1.36k| idx_s += 4; 883| 1.36k| idx_l += index4; 884| 1.36k| } 885| 743| if ((idx_s + 2 <= size_s) && (idx_l < size_l)) { ------------------ | Branch (885:9): [True: 383, False: 360] | Branch (885:34): [True: 381, False: 2] ------------------ 886| 381| uint16_t target1 = small[idx_s]; 887| 381| uint16_t target2 = small[idx_s + 1]; 888| 381| binarySearch2(large + idx_l, (int32_t)(size_l - idx_l), target1, 889| 381| target2, &index1, &index2); 890| 381| if ((index1 + idx_l < size_l) && (large[idx_l + index1] == target1)) { ------------------ | Branch (890:13): [True: 375, False: 6] | Branch (890:42): [True: 239, False: 136] ------------------ 891| 239| buffer[pos++] = target1; 892| 239| } 893| 381| if ((index2 + idx_l < size_l) && (large[idx_l + index2] == target2)) { ------------------ | Branch (893:13): [True: 365, False: 16] | Branch (893:42): [True: 241, False: 124] ------------------ 894| 241| buffer[pos++] = target2; 895| 241| } 896| 381| idx_s += 2; 897| 381| idx_l += index2; 898| 381| } 899| 743| if ((idx_s < size_s) && (idx_l < size_l)) { ------------------ | Branch (899:9): [True: 356, False: 387] | Branch (899:29): [True: 353, False: 3] ------------------ 900| 353| uint16_t val_s = small[idx_s]; 901| 353| int32_t index = 902| 353| binarySearch(large + idx_l, (int32_t)(size_l - idx_l), val_s); 903| 353| if (index >= 0) buffer[pos++] = val_s; ------------------ | Branch (903:13): [True: 247, False: 106] ------------------ 904| 353| } 905| 743| return (int32_t)pos; 906| 743|} intersect_skewed_uint16_cardinality: 912| 2.15k| size_t size_l) { 913| 2.15k| size_t pos = 0, idx_l = 0, idx_s = 0; 914| | 915| 2.15k| if (0 == size_s) { ------------------ | Branch (915:9): [True: 0, False: 2.15k] ------------------ 916| 0| return 0; 917| 0| } 918| | 919| 2.15k| uint16_t val_l = large[idx_l], val_s = small[idx_s]; 920| | 921| 30.5k| while (true) { ------------------ | Branch (921:12): [True: 30.5k, Folded] ------------------ 922| 30.5k| if (val_l < val_s) { ------------------ | Branch (922:13): [True: 6.08k, False: 24.4k] ------------------ 923| 6.08k| idx_l = advanceUntil(large, (int32_t)idx_l, (int32_t)size_l, val_s); 924| 6.08k| if (idx_l == size_l) break; ------------------ | Branch (924:17): [True: 140, False: 5.94k] ------------------ 925| 5.94k| val_l = large[idx_l]; 926| 24.4k| } else if (val_s < val_l) { ------------------ | Branch (926:20): [True: 10.4k, False: 13.9k] ------------------ 927| 10.4k| idx_s++; 928| 10.4k| if (idx_s == size_s) break; ------------------ | Branch (928:17): [True: 840, False: 9.65k] ------------------ 929| 9.65k| val_s = small[idx_s]; 930| 13.9k| } else { 931| 13.9k| pos++; 932| 13.9k| idx_s++; 933| 13.9k| if (idx_s == size_s) break; ------------------ | Branch (933:17): [True: 1.10k, False: 12.8k] ------------------ 934| 12.8k| val_s = small[idx_s]; 935| 12.8k| idx_l = advanceUntil(large, (int32_t)idx_l, (int32_t)size_l, val_s); 936| 12.8k| if (idx_l == size_l) break; ------------------ | Branch (936:17): [True: 72, False: 12.7k] ------------------ 937| 12.7k| val_l = large[idx_l]; 938| 12.7k| } 939| 30.5k| } 940| | 941| 2.15k| return (int32_t)pos; 942| 2.15k|} intersect_skewed_uint16_nonempty: 945| 538| const uint16_t *large, size_t size_l) { 946| 538| size_t idx_l = 0, idx_s = 0; 947| | 948| 538| if (0 == size_s) { ------------------ | Branch (948:9): [True: 0, False: 538] ------------------ 949| 0| return false; 950| 0| } 951| | 952| 538| uint16_t val_l = large[idx_l], val_s = small[idx_s]; 953| | 954| 2.06k| while (true) { ------------------ | Branch (954:12): [True: 2.06k, Folded] ------------------ 955| 2.06k| if (val_l < val_s) { ------------------ | Branch (955:13): [True: 793, False: 1.26k] ------------------ 956| 793| idx_l = advanceUntil(large, (int32_t)idx_l, (int32_t)size_l, val_s); 957| 793| if (idx_l == size_l) break; ------------------ | Branch (957:17): [True: 34, False: 759] ------------------ 958| 759| val_l = large[idx_l]; 959| 1.26k| } else if (val_s < val_l) { ------------------ | Branch (959:20): [True: 845, False: 423] ------------------ 960| 845| idx_s++; 961| 845| if (idx_s == size_s) break; ------------------ | Branch (961:17): [True: 81, False: 764] ------------------ 962| 764| val_s = small[idx_s]; 963| 764| } else { 964| 423| return true; 965| 423| } 966| 2.06k| } 967| | 968| 115| return false; 969| 538|} intersect_uint16_nonempty: 1025| 1.76k| const uint16_t *B, const size_t lenB) { 1026| 1.76k| if (lenA == 0 || lenB == 0) return 0; ------------------ | Branch (1026:9): [True: 0, False: 1.76k] | Branch (1026:22): [True: 0, False: 1.76k] ------------------ 1027| 1.76k| const uint16_t *endA = A + lenA; 1028| 1.76k| const uint16_t *endB = B + lenB; 1029| | 1030| 1.76k| while (1) { ------------------ | Branch (1030:12): [True: 1.76k, Folded] ------------------ 1031| 9.93k| while (*A < *B) { ------------------ | Branch (1031:16): [True: 7.74k, False: 2.19k] ------------------ 1032| 8.21k| SKIP_FIRST_COMPARE: 1033| 8.21k| if (++A == endA) return false; ------------------ | Branch (1033:17): [True: 44, False: 8.17k] ------------------ 1034| 8.21k| } 1035| 4.55k| while (*A > *B) { ------------------ | Branch (1035:16): [True: 2.37k, False: 2.18k] ------------------ 1036| 2.37k| if (++B == endB) return false; ------------------ | Branch (1036:17): [True: 8, False: 2.36k] ------------------ 1037| 2.37k| } 1038| 2.18k| if (*A == *B) { ------------------ | Branch (1038:13): [True: 1.71k, False: 469] ------------------ 1039| 1.71k| return true; 1040| 1.71k| } else { 1041| 469| goto SKIP_FIRST_COMPARE; 1042| 469| } 1043| 2.18k| } 1044| 0| return false; // NOTREACHED 1045| 1.76k|} union_uint16: 1105| 4.59k| size_t size_2, uint16_t *buffer) { 1106| 4.59k| size_t pos = 0, idx_1 = 0, idx_2 = 0; 1107| | 1108| 4.59k| if (0 == size_2) { ------------------ | Branch (1108:9): [True: 60, False: 4.53k] ------------------ 1109| 60| memmove(buffer, set_1, size_1 * sizeof(uint16_t)); 1110| 60| return size_1; 1111| 60| } 1112| 4.53k| if (0 == size_1) { ------------------ | Branch (1112:9): [True: 0, False: 4.53k] ------------------ 1113| 0| memmove(buffer, set_2, size_2 * sizeof(uint16_t)); 1114| 0| return size_2; 1115| 0| } 1116| | 1117| 4.53k| uint16_t val_1 = set_1[idx_1], val_2 = set_2[idx_2]; 1118| | 1119| 209k| while (true) { ------------------ | Branch (1119:12): [True: 209k, Folded] ------------------ 1120| 209k| if (val_1 < val_2) { ------------------ | Branch (1120:13): [True: 7.94k, False: 201k] ------------------ 1121| 7.94k| buffer[pos++] = val_1; 1122| 7.94k| ++idx_1; 1123| 7.94k| if (idx_1 >= size_1) break; ------------------ | Branch (1123:17): [True: 391, False: 7.55k] ------------------ 1124| 7.55k| val_1 = set_1[idx_1]; 1125| 201k| } else if (val_2 < val_1) { ------------------ | Branch (1125:20): [True: 169k, False: 31.3k] ------------------ 1126| 169k| buffer[pos++] = val_2; 1127| 169k| ++idx_2; 1128| 169k| if (idx_2 >= size_2) break; ------------------ | Branch (1128:17): [True: 201, False: 169k] ------------------ 1129| 169k| val_2 = set_2[idx_2]; 1130| 169k| } else { 1131| 31.3k| buffer[pos++] = val_1; 1132| 31.3k| ++idx_1; 1133| 31.3k| ++idx_2; 1134| 31.3k| if (idx_1 >= size_1 || idx_2 >= size_2) break; ------------------ | Branch (1134:17): [True: 3.81k, False: 27.5k] | Branch (1134:36): [True: 121, False: 27.4k] ------------------ 1135| 27.4k| val_1 = set_1[idx_1]; 1136| 27.4k| val_2 = set_2[idx_2]; 1137| 27.4k| } 1138| 209k| } 1139| | 1140| 4.53k| if (idx_1 < size_1) { ------------------ | Branch (1140:9): [True: 322, False: 4.20k] ------------------ 1141| 322| const size_t n_elems = size_1 - idx_1; 1142| 322| memmove(buffer + pos, set_1 + idx_1, n_elems * sizeof(uint16_t)); 1143| 322| pos += n_elems; 1144| 4.20k| } else if (idx_2 < size_2) { ------------------ | Branch (1144:16): [True: 1.76k, False: 2.44k] ------------------ 1145| 1.76k| const size_t n_elems = size_2 - idx_2; 1146| 1.76k| memmove(buffer + pos, set_2 + idx_2, n_elems * sizeof(uint16_t)); 1147| 1.76k| pos += n_elems; 1148| 1.76k| } 1149| | 1150| 4.53k| return pos; 1151| 4.53k|} difference_uint16: 1154| 3.44k| int length2, uint16_t *a_out) { 1155| 3.44k| int out_card = 0; 1156| 3.44k| int k1 = 0, k2 = 0; 1157| 3.44k| if (length1 == 0) return 0; ------------------ | Branch (1157:9): [True: 0, False: 3.44k] ------------------ 1158| 3.44k| if (length2 == 0) { ------------------ | Branch (1158:9): [True: 0, False: 3.44k] ------------------ 1159| 0| if (a1 != a_out) memcpy(a_out, a1, sizeof(uint16_t) * length1); ------------------ | Branch (1159:13): [True: 0, False: 0] ------------------ 1160| 0| return length1; 1161| 0| } 1162| 3.44k| uint16_t s1 = a1[k1]; 1163| 3.44k| uint16_t s2 = a2[k2]; 1164| 421k| while (true) { ------------------ | Branch (1164:12): [True: 421k, Folded] ------------------ 1165| 421k| if (s1 < s2) { ------------------ | Branch (1165:13): [True: 207k, False: 214k] ------------------ 1166| 207k| a_out[out_card++] = s1; 1167| 207k| ++k1; 1168| 207k| if (k1 >= length1) { ------------------ | Branch (1168:17): [True: 0, False: 207k] ------------------ 1169| 0| break; 1170| 0| } 1171| 207k| s1 = a1[k1]; 1172| 214k| } else if (s1 == s2) { ------------------ | Branch (1172:20): [True: 214k, False: 0] ------------------ 1173| 214k| ++k1; 1174| 214k| ++k2; 1175| 214k| if (k1 >= length1) { ------------------ | Branch (1175:17): [True: 2.23k, False: 212k] ------------------ 1176| 2.23k| break; 1177| 2.23k| } 1178| 212k| if (k2 >= length2) { ------------------ | Branch (1178:17): [True: 1.20k, False: 210k] ------------------ 1179| 1.20k| memmove(a_out + out_card, a1 + k1, 1180| 1.20k| sizeof(uint16_t) * (length1 - k1)); 1181| 1.20k| return out_card + length1 - k1; 1182| 1.20k| } 1183| 210k| s1 = a1[k1]; 1184| 210k| s2 = a2[k2]; 1185| 210k| } else { // if (val1>val2) 1186| 0| ++k2; 1187| 0| if (k2 >= length2) { ------------------ | Branch (1187:17): [True: 0, False: 0] ------------------ 1188| 0| memmove(a_out + out_card, a1 + k1, 1189| 0| sizeof(uint16_t) * (length1 - k1)); 1190| 0| return out_card + length1 - k1; 1191| 0| } 1192| 0| s2 = a2[k2]; 1193| 0| } 1194| 421k| } 1195| 2.23k| return out_card; 1196| 3.44k|} xor_uint16: 1199| 7.70k| const uint16_t *array_2, int32_t card_2, uint16_t *out) { 1200| 7.70k| int32_t pos1 = 0, pos2 = 0, pos_out = 0; 1201| 219k| while (pos1 < card_1 && pos2 < card_2) { ------------------ | Branch (1201:12): [True: 212k, False: 7.12k] | Branch (1201:29): [True: 212k, False: 576] ------------------ 1202| 212k| const uint16_t v1 = array_1[pos1]; 1203| 212k| const uint16_t v2 = array_2[pos2]; 1204| 212k| if (v1 == v2) { ------------------ | Branch (1204:13): [True: 38.4k, False: 173k] ------------------ 1205| 38.4k| ++pos1; 1206| 38.4k| ++pos2; 1207| 38.4k| continue; 1208| 38.4k| } 1209| 173k| if (v1 < v2) { ------------------ | Branch (1209:13): [True: 30.3k, False: 143k] ------------------ 1210| 30.3k| out[pos_out++] = v1; 1211| 30.3k| ++pos1; 1212| 143k| } else { 1213| 143k| out[pos_out++] = v2; 1214| 143k| ++pos2; 1215| 143k| } 1216| 173k| } 1217| 7.70k| if (pos1 < card_1) { ------------------ | Branch (1217:9): [True: 576, False: 7.12k] ------------------ 1218| 576| const size_t n_elems = card_1 - pos1; 1219| 576| memcpy(out + pos_out, array_1 + pos1, n_elems * sizeof(uint16_t)); 1220| 576| pos_out += (int32_t)n_elems; 1221| 7.12k| } else if (pos2 < card_2) { ------------------ | Branch (1221:16): [True: 2.27k, False: 4.85k] ------------------ 1222| 2.27k| const size_t n_elems = card_2 - pos2; 1223| 2.27k| memcpy(out + pos_out, array_2 + pos2, n_elems * sizeof(uint16_t)); 1224| 2.27k| pos_out += (int32_t)n_elems; 1225| 2.27k| } 1226| 7.70k| return pos_out; 1227| 7.70k|} union_vector16: 1664| 4.59k| uint16_t *output) { 1665| 4.59k| if ((length1 < 8) || (length2 < 8)) { ------------------ | Branch (1665:9): [True: 1.01k, False: 3.57k] | Branch (1665:26): [True: 0, False: 3.57k] ------------------ 1666| 1.01k| return (uint32_t)union_uint16(array1, length1, array2, length2, output); 1667| 1.01k| } 1668| 3.57k| __m128i vA, vB, V, vecMin, vecMax; 1669| 3.57k| __m128i laststore; 1670| 3.57k| uint16_t *initoutput = output; 1671| 3.57k| uint32_t len1 = length1 / 8; 1672| 3.57k| uint32_t len2 = length2 / 8; 1673| 3.57k| uint32_t pos1 = 0; 1674| 3.57k| uint32_t pos2 = 0; 1675| | // we start the machine 1676| 3.57k| vA = _mm_lddqu_si128((const __m128i *)array1 + pos1); 1677| 3.57k| pos1++; 1678| 3.57k| vB = _mm_lddqu_si128((const __m128i *)array2 + pos2); 1679| 3.57k| pos2++; 1680| 3.57k| sse_merge(&vA, &vB, &vecMin, &vecMax); 1681| 3.57k| laststore = _mm_set1_epi16(-1); 1682| 3.57k| output += store_unique(laststore, vecMin, output); 1683| 3.57k| laststore = vecMin; 1684| 3.57k| if ((pos1 < len1) && (pos2 < len2)) { ------------------ | Branch (1684:9): [True: 3.01k, False: 565] | Branch (1684:26): [True: 3.01k, False: 0] ------------------ 1685| 3.01k| uint16_t curA, curB; 1686| 3.01k| curA = array1[8 * pos1]; 1687| 3.01k| curB = array2[8 * pos2]; 1688| 101k| while (true) { ------------------ | Branch (1688:16): [True: 101k, Folded] ------------------ 1689| 101k| if (curA <= curB) { ------------------ | Branch (1689:17): [True: 43.0k, False: 58.0k] ------------------ 1690| 43.0k| V = _mm_lddqu_si128((const __m128i *)array1 + pos1); 1691| 43.0k| pos1++; 1692| 43.0k| if (pos1 < len1) { ------------------ | Branch (1692:21): [True: 40.7k, False: 2.30k] ------------------ 1693| 40.7k| curA = array1[8 * pos1]; 1694| 40.7k| } else { 1695| 2.30k| break; 1696| 2.30k| } 1697| 58.0k| } else { 1698| 58.0k| V = _mm_lddqu_si128((const __m128i *)array2 + pos2); 1699| 58.0k| pos2++; 1700| 58.0k| if (pos2 < len2) { ------------------ | Branch (1700:21): [True: 57.3k, False: 711] ------------------ 1701| 57.3k| curB = array2[8 * pos2]; 1702| 57.3k| } else { 1703| 711| break; 1704| 711| } 1705| 58.0k| } 1706| 98.0k| sse_merge(&V, &vecMax, &vecMin, &vecMax); 1707| 98.0k| output += store_unique(laststore, vecMin, output); 1708| 98.0k| laststore = vecMin; 1709| 98.0k| } 1710| 3.01k| sse_merge(&V, &vecMax, &vecMin, &vecMax); 1711| 3.01k| output += store_unique(laststore, vecMin, output); 1712| 3.01k| laststore = vecMin; 1713| 3.01k| } 1714| | // we finish the rest off using a scalar algorithm 1715| | // could be improved? 1716| | // 1717| | // copy the small end on a tmp buffer 1718| 3.57k| uint32_t len = (uint32_t)(output - initoutput); 1719| 3.57k| uint16_t buffer[16]; 1720| 3.57k| uint32_t leftoversize = store_unique(laststore, vecMax, buffer); 1721| 3.57k| if (pos1 == len1) { ------------------ | Branch (1721:9): [True: 2.86k, False: 711] ------------------ 1722| 2.86k| memcpy(buffer + leftoversize, array1 + 8 * pos1, 1723| 2.86k| (length1 - 8 * len1) * sizeof(uint16_t)); 1724| 2.86k| leftoversize += length1 - 8 * len1; 1725| 2.86k| sort_uint16_short(buffer, leftoversize); 1726| | 1727| 2.86k| leftoversize = unique(buffer, leftoversize); 1728| 2.86k| len += (uint32_t)union_uint16(buffer, leftoversize, array2 + 8 * pos2, 1729| 2.86k| length2 - 8 * pos2, output); 1730| 2.86k| } else { 1731| 711| memcpy(buffer + leftoversize, array2 + 8 * pos2, 1732| 711| (length2 - 8 * len2) * sizeof(uint16_t)); 1733| 711| leftoversize += length2 - 8 * len2; 1734| 711| sort_uint16_short(buffer, leftoversize); 1735| 711| leftoversize = unique(buffer, leftoversize); 1736| 711| len += (uint32_t)union_uint16(buffer, leftoversize, array1 + 8 * pos1, 1737| 711| length1 - 8 * pos1, output); 1738| 711| } 1739| 3.57k| return len; 1740| 4.59k|} xor_vector16: 1975| 7.71k| uint16_t *output) { 1976| 7.71k| if ((length1 < 8) || (length2 < 8)) { ------------------ | Branch (1976:9): [True: 3.03k, False: 4.67k] | Branch (1976:26): [True: 348, False: 4.33k] ------------------ 1977| 3.38k| return xor_uint16(array1, length1, array2, length2, output); 1978| 3.38k| } 1979| 4.33k| __m128i vA, vB, V, vecMin, vecMax; 1980| 4.33k| __m128i laststore; 1981| 4.33k| uint16_t *initoutput = output; 1982| 4.33k| uint32_t len1 = length1 / 8; 1983| 4.33k| uint32_t len2 = length2 / 8; 1984| 4.33k| uint32_t pos1 = 0; 1985| 4.33k| uint32_t pos2 = 0; 1986| | // we start the machine 1987| 4.33k| vA = _mm_lddqu_si128((const __m128i *)array1 + pos1); 1988| 4.33k| pos1++; 1989| 4.33k| vB = _mm_lddqu_si128((const __m128i *)array2 + pos2); 1990| 4.33k| pos2++; 1991| 4.33k| sse_merge(&vA, &vB, &vecMin, &vecMax); 1992| 4.33k| laststore = _mm_set1_epi16(-1); 1993| 4.33k| uint16_t buffer[17]; 1994| 4.33k| output += store_unique_xor(laststore, vecMin, output); 1995| | 1996| 4.33k| laststore = vecMin; 1997| 4.33k| if ((pos1 < len1) && (pos2 < len2)) { ------------------ | Branch (1997:9): [True: 3.97k, False: 355] | Branch (1997:26): [True: 3.85k, False: 120] ------------------ 1998| 3.85k| uint16_t curA, curB; 1999| 3.85k| curA = array1[8 * pos1]; 2000| 3.85k| curB = array2[8 * pos2]; 2001| 181k| while (true) { ------------------ | Branch (2001:16): [True: 181k, Folded] ------------------ 2002| 181k| if (curA <= curB) { ------------------ | Branch (2002:17): [True: 86.1k, False: 94.8k] ------------------ 2003| 86.1k| V = _mm_lddqu_si128((const __m128i *)array1 + pos1); 2004| 86.1k| pos1++; 2005| 86.1k| if (pos1 < len1) { ------------------ | Branch (2005:21): [True: 83.6k, False: 2.53k] ------------------ 2006| 83.6k| curA = array1[8 * pos1]; 2007| 83.6k| } else { 2008| 2.53k| break; 2009| 2.53k| } 2010| 94.8k| } else { 2011| 94.8k| V = _mm_lddqu_si128((const __m128i *)array2 + pos2); 2012| 94.8k| pos2++; 2013| 94.8k| if (pos2 < len2) { ------------------ | Branch (2013:21): [True: 93.5k, False: 1.31k] ------------------ 2014| 93.5k| curB = array2[8 * pos2]; 2015| 93.5k| } else { 2016| 1.31k| break; 2017| 1.31k| } 2018| 94.8k| } 2019| 177k| sse_merge(&V, &vecMax, &vecMin, &vecMax); 2020| | // conditionally stores the last value of laststore as well as all 2021| | // but the 2022| | // last value of vecMin 2023| 177k| output += store_unique_xor(laststore, vecMin, output); 2024| 177k| laststore = vecMin; 2025| 177k| } 2026| 3.85k| sse_merge(&V, &vecMax, &vecMin, &vecMax); 2027| | // conditionally stores the last value of laststore as well as all but 2028| | // the 2029| | // last value of vecMin 2030| 3.85k| output += store_unique_xor(laststore, vecMin, output); 2031| 3.85k| laststore = vecMin; 2032| 3.85k| } 2033| 4.33k| uint32_t len = (uint32_t)(output - initoutput); 2034| | 2035| | // we finish the rest off using a scalar algorithm 2036| | // could be improved? 2037| | // conditionally stores the last value of laststore as well as all but the 2038| | // last value of vecMax, 2039| | // we store to "buffer" 2040| 4.33k| int leftoversize = store_unique_xor(laststore, vecMax, buffer); 2041| 4.33k| uint16_t vec7 = (uint16_t)_mm_extract_epi16(vecMax, 7); 2042| 4.33k| uint16_t vec6 = (uint16_t)_mm_extract_epi16(vecMax, 6); 2043| 4.33k| if (vec7 != vec6) buffer[leftoversize++] = vec7; ------------------ | Branch (2043:9): [True: 4.26k, False: 65] ------------------ 2044| 4.33k| if (pos1 == len1) { ------------------ | Branch (2044:9): [True: 2.89k, False: 1.43k] ------------------ 2045| 2.89k| memcpy(buffer + leftoversize, array1 + 8 * pos1, 2046| 2.89k| (length1 - 8 * len1) * sizeof(uint16_t)); 2047| 2.89k| leftoversize += length1 - 8 * len1; 2048| 2.89k| if (leftoversize == 0) { // trivial case ------------------ | Branch (2048:13): [True: 8, False: 2.88k] ------------------ 2049| 8| memcpy(output, array2 + 8 * pos2, 2050| 8| (length2 - 8 * pos2) * sizeof(uint16_t)); 2051| 8| len += (length2 - 8 * pos2); 2052| 2.88k| } else { 2053| 2.88k| sort_uint16_short(buffer, leftoversize); 2054| 2.88k| leftoversize = unique_xor(buffer, leftoversize); 2055| 2.88k| len += xor_uint16(buffer, leftoversize, array2 + 8 * pos2, 2056| 2.88k| length2 - 8 * pos2, output); 2057| 2.88k| } 2058| 2.89k| } else { 2059| 1.43k| memcpy(buffer + leftoversize, array2 + 8 * pos2, 2060| 1.43k| (length2 - 8 * len2) * sizeof(uint16_t)); 2061| 1.43k| leftoversize += length2 - 8 * len2; 2062| 1.43k| if (leftoversize == 0) { // trivial case ------------------ | Branch (2062:13): [True: 4, False: 1.43k] ------------------ 2063| 4| memcpy(output, array1 + 8 * pos1, 2064| 4| (length1 - 8 * pos1) * sizeof(uint16_t)); 2065| 4| len += (length1 - 8 * pos1); 2066| 1.43k| } else { 2067| 1.43k| sort_uint16_short(buffer, leftoversize); 2068| 1.43k| leftoversize = unique_xor(buffer, leftoversize); 2069| 1.43k| len += xor_uint16(buffer, leftoversize, array1 + 8 * pos1, 2070| 1.43k| length1 - 8 * pos1, output); 2071| 1.43k| } 2072| 1.43k| } 2073| 4.33k| return len; 2074| 7.71k|} fast_union_uint16: 2177| 4.59k| uint16_t *buffer) { 2178| 4.59k|#if CROARING_IS_X64 2179| 4.59k| const unsigned support = (unsigned)croaring_hardware_support(); 2180| 4.59k|#if CROARING_COMPILER_SUPPORTS_AVX512 2181| 4.59k| if (support & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (2181:9): [True: 0, False: 4.59k] ------------------ 2182| | // compute union with smallest array first 2183| 0| if (size_1 < size_2) { ------------------ | Branch (2183:13): [True: 0, False: 0] ------------------ 2184| 0| return avx512_union_uint16(set_1, (uint32_t)size_1, set_2, 2185| 0| (uint32_t)size_2, buffer); 2186| 0| } else { 2187| 0| return avx512_union_uint16(set_2, (uint32_t)size_2, set_1, 2188| 0| (uint32_t)size_1, buffer); 2189| 0| } 2190| 0| } 2191| 4.59k|#endif // CROARING_COMPILER_SUPPORTS_AVX512 2192| 4.59k| if (support & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (2192:9): [True: 4.59k, False: 0] ------------------ 2193| | // compute union with smallest array first 2194| 4.59k| if (size_1 < size_2) { ------------------ | Branch (2194:13): [True: 2.30k, False: 2.28k] ------------------ 2195| 2.30k| return union_vector16(set_1, (uint32_t)size_1, set_2, 2196| 2.30k| (uint32_t)size_2, buffer); 2197| 2.30k| } else { 2198| 2.28k| return union_vector16(set_2, (uint32_t)size_2, set_1, 2199| 2.28k| (uint32_t)size_1, buffer); 2200| 2.28k| } 2201| 4.59k| } else { 2202| | // compute union with smallest array first 2203| 0| if (size_1 < size_2) { ------------------ | Branch (2203:13): [True: 0, False: 0] ------------------ 2204| 0| return union_uint16(set_1, size_1, set_2, size_2, buffer); 2205| 0| } else { 2206| 0| return union_uint16(set_2, size_2, set_1, size_1, buffer); 2207| 0| } 2208| 0| } 2209| |#else 2210| | // compute union with smallest array first 2211| | if (size_1 < size_2) { 2212| | return union_uint16(set_1, size_1, set_2, size_2, buffer); 2213| | } else { 2214| | return union_uint16(set_2, size_2, set_1, size_1, buffer); 2215| | } 2216| |#endif 2217| 4.59k|} memequals: 2321| 6.99k|bool memequals(const void *s1, const void *s2, size_t n) { 2322| 6.99k| if (n == 0) { ------------------ | Branch (2322:9): [True: 0, False: 6.99k] ------------------ 2323| 0| return true; 2324| 0| } 2325| 6.99k|#if CROARING_IS_X64 2326| 6.99k| int support = croaring_hardware_support(); 2327| 6.99k|#if CROARING_COMPILER_SUPPORTS_AVX512 2328| 6.99k| if (support & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (2328:9): [True: 0, False: 6.99k] ------------------ 2329| 0| return _avx512_memequals(s1, s2, n); 2330| 0| } else 2331| 6.99k|#endif // CROARING_COMPILER_SUPPORTS_AVX512 2332| 6.99k| if (support & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (2332:13): [True: 6.99k, False: 0] ------------------ 2333| 6.99k| return _avx2_memequals(s1, s2, n); 2334| 6.99k| } else { 2335| 0| return memcmp(s1, s2, n) == 0; 2336| 0| } 2337| |#else 2338| | return memcmp(s1, s2, n) == 0; 2339| |#endif 2340| 6.99k|} array_util.c:binarySearch4: 804| 1.36k| int32_t *index4) { 805| 1.36k| const uint16_t *base1 = array; 806| 1.36k| const uint16_t *base2 = array; 807| 1.36k| const uint16_t *base3 = array; 808| 1.36k| const uint16_t *base4 = array; 809| 1.36k| if (n == 0) return; ------------------ | Branch (809:9): [True: 0, False: 1.36k] ------------------ 810| 17.0k| while (n > 1) { ------------------ | Branch (810:12): [True: 15.7k, False: 1.36k] ------------------ 811| 15.7k| int32_t half = n >> 1; 812| 15.7k| base1 = (base1[half] < target1) ? &base1[half] : base1; ------------------ | Branch (812:17): [True: 1.15k, False: 14.5k] ------------------ 813| 15.7k| base2 = (base2[half] < target2) ? &base2[half] : base2; ------------------ | Branch (813:17): [True: 2.21k, False: 13.5k] ------------------ 814| 15.7k| base3 = (base3[half] < target3) ? &base3[half] : base3; ------------------ | Branch (814:17): [True: 2.76k, False: 12.9k] ------------------ 815| 15.7k| base4 = (base4[half] < target4) ? &base4[half] : base4; ------------------ | Branch (815:17): [True: 3.34k, False: 12.3k] ------------------ 816| 15.7k| n -= half; 817| 15.7k| } 818| 1.36k| *index1 = (int32_t)((*base1 < target1) + base1 - array); 819| 1.36k| *index2 = (int32_t)((*base2 < target2) + base2 - array); 820| 1.36k| *index3 = (int32_t)((*base3 < target3) + base3 - array); 821| 1.36k| *index4 = (int32_t)((*base4 < target4) + base4 - array); 822| 1.36k|} array_util.c:binarySearch2: 833| 381| uint16_t target2, int32_t *index1, int32_t *index2) { 834| 381| const uint16_t *base1 = array; 835| 381| const uint16_t *base2 = array; 836| 381| if (n == 0) return; ------------------ | Branch (836:9): [True: 0, False: 381] ------------------ 837| 3.57k| while (n > 1) { ------------------ | Branch (837:12): [True: 3.19k, False: 381] ------------------ 838| 3.19k| int32_t half = n >> 1; 839| 3.19k| base1 = (base1[half] < target1) ? &base1[half] : base1; ------------------ | Branch (839:17): [True: 792, False: 2.40k] ------------------ 840| 3.19k| base2 = (base2[half] < target2) ? &base2[half] : base2; ------------------ | Branch (840:17): [True: 1.32k, False: 1.87k] ------------------ 841| 3.19k| n -= half; 842| 3.19k| } 843| 381| *index1 = (int32_t)((*base1 < target1) + base1 - array); 844| 381| *index2 = (int32_t)((*base2 < target2) + base2 - array); 845| 381|} array_util.c:sse_merge: 1244| 289k| __m128i *vecMin, __m128i *vecMax) { // output 1245| 289k| __m128i vecTmp; 1246| 289k| vecTmp = _mm_min_epu16(*vInput1, *vInput2); 1247| 289k| *vecMax = _mm_max_epu16(*vInput1, *vInput2); 1248| 289k| vecTmp = _mm_alignr_epi8(vecTmp, vecTmp, 2); 1249| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1250| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1251| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1252| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1253| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1254| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1255| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1256| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1257| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1258| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1259| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1260| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1261| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1262| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1263| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1264| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1265| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1266| 289k| vecTmp = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1267| 289k| *vecMin = _mm_min_epu16(vecTmp, *vecMax); 1268| 289k| *vecMax = _mm_max_epu16(vecTmp, *vecMax); 1269| | *vecMin = _mm_alignr_epi8(*vecMin, *vecMin, 2); 1270| 289k|} array_util.c:store_unique: 1619| 108k|static inline int store_unique(__m128i old, __m128i newval, uint16_t *output) { 1620| 108k| __m128i vecTmp = _mm_alignr_epi8(newval, old, 16 - 2); 1621| | // lots of high latency instructions follow (optimize?) 1622| 108k| int M = _mm_movemask_epi8( 1623| 108k| _mm_packs_epi16(_mm_cmpeq_epi16(vecTmp, newval), _mm_setzero_si128())); 1624| 108k| int numberofnewvalues = 8 - _mm_popcnt_u32(M); 1625| 108k| __m128i key = _mm_lddqu_si128((const __m128i *)uniqshuf + M); 1626| 108k| __m128i val = _mm_shuffle_epi8(newval, key); 1627| 108k| _mm_storeu_si128((__m128i *)output, val); 1628| 108k| return numberofnewvalues; 1629| 108k|} array_util.c:sort_uint16_short: 1647| 7.89k|static inline void sort_uint16_short(uint16_t *z, uint32_t n) { 1648| 79.1k| for (uint32_t i = 1; i < n; i++) { ------------------ | Branch (1648:26): [True: 71.2k, False: 7.89k] ------------------ 1649| 71.2k| uint16_t v = z[i]; 1650| 71.2k| uint32_t j = i; 1651| 74.4k| while (j > 0 && z[j - 1] > v) { ------------------ | Branch (1651:16): [True: 74.3k, False: 28] | Branch (1651:25): [True: 3.20k, False: 71.1k] ------------------ 1652| 3.20k| z[j] = z[j - 1]; 1653| 3.20k| j--; 1654| 3.20k| } 1655| 71.2k| z[j] = v; 1656| 71.2k| } 1657| 7.89k|} array_util.c:unique: 1634| 3.57k|static inline uint32_t unique(uint16_t *out, uint32_t len) { 1635| 3.57k| uint32_t pos = 1; 1636| 36.5k| for (uint32_t i = 1; i < len; ++i) { ------------------ | Branch (1636:26): [True: 32.9k, False: 3.57k] ------------------ 1637| 32.9k| if (out[i] != out[i - 1]) { ------------------ | Branch (1637:13): [True: 32.8k, False: 133] ------------------ 1638| 32.8k| out[pos++] = out[i]; 1639| 32.8k| } 1640| 32.9k| } 1641| 3.57k| return pos; 1642| 3.57k|} array_util.c:store_unique_xor: 1943| 189k| uint16_t *output) { 1944| 189k| __m128i vecTmp1 = _mm_alignr_epi8(newval, old, 16 - 4); 1945| | __m128i vecTmp2 = _mm_alignr_epi8(newval, old, 16 - 2); 1946| 189k| __m128i equalleft = _mm_cmpeq_epi16(vecTmp2, vecTmp1); 1947| 189k| __m128i equalright = _mm_cmpeq_epi16(vecTmp2, newval); 1948| 189k| __m128i equalleftoright = _mm_or_si128(equalleft, equalright); 1949| 189k| int M = _mm_movemask_epi8( 1950| 189k| _mm_packs_epi16(equalleftoright, _mm_setzero_si128())); 1951| 189k| int numberofnewvalues = 8 - _mm_popcnt_u32(M); 1952| 189k| __m128i key = _mm_lddqu_si128((const __m128i *)uniqshuf + M); 1953| 189k| __m128i val = _mm_shuffle_epi8(vecTmp2, key); 1954| 189k| _mm_storeu_si128((__m128i *)output, val); 1955| 189k| return numberofnewvalues; 1956| 189k|} array_util.c:unique_xor: 1961| 4.31k|static inline uint32_t unique_xor(uint16_t *out, uint32_t len) { 1962| 4.31k| uint32_t pos = 1; 1963| 42.5k| for (uint32_t i = 1; i < len; ++i) { ------------------ | Branch (1963:26): [True: 38.2k, False: 4.31k] ------------------ 1964| 38.2k| if (out[i] != out[i - 1]) { ------------------ | Branch (1964:13): [True: 37.9k, False: 294] ------------------ 1965| 37.9k| out[pos++] = out[i]; 1966| 37.9k| } else 1967| 294| pos--; // if it is identical to previous, delete it 1968| 38.2k| } 1969| 4.31k| return pos; 1970| 4.31k|} array_util.c:_avx2_memequals: 2279| 6.99k|static inline bool _avx2_memequals(const void *s1, const void *s2, size_t n) { 2280| 6.99k| const uint8_t *ptr1 = (const uint8_t *)s1; 2281| 6.99k| const uint8_t *ptr2 = (const uint8_t *)s2; 2282| 6.99k| const uint8_t *end1 = ptr1 + n; 2283| 6.99k| const uint8_t *end8 = ptr1 + n / 8 * 8; 2284| 6.99k| const uint8_t *end32 = ptr1 + n / 32 * 32; 2285| | 2286| 134k| while (ptr1 < end32) { ------------------ | Branch (2286:12): [True: 127k, False: 6.93k] ------------------ 2287| 127k| __m256i r1 = _mm256_loadu_si256((const __m256i *)ptr1); 2288| 127k| __m256i r2 = _mm256_loadu_si256((const __m256i *)ptr2); 2289| 127k| int mask = _mm256_movemask_epi8(_mm256_cmpeq_epi8(r1, r2)); 2290| 127k| if ((uint32_t)mask != UINT32_MAX) { ------------------ | Branch (2290:13): [True: 63, False: 127k] ------------------ 2291| 63| return false; 2292| 63| } 2293| 127k| ptr1 += 32; 2294| 127k| ptr2 += 32; 2295| 127k| } 2296| | 2297| 12.7k| while (ptr1 < end8) { ------------------ | Branch (2297:12): [True: 5.92k, False: 6.85k] ------------------ 2298| 5.92k| uint64_t v1, v2; 2299| 5.92k| memcpy(&v1, ptr1, sizeof(uint64_t)); 2300| 5.92k| memcpy(&v2, ptr2, sizeof(uint64_t)); 2301| 5.92k| if (v1 != v2) { ------------------ | Branch (2301:13): [True: 77, False: 5.84k] ------------------ 2302| 77| return false; 2303| 77| } 2304| 5.84k| ptr1 += 8; 2305| 5.84k| ptr2 += 8; 2306| 5.84k| } 2307| | 2308| 24.6k| while (ptr1 < end1) { ------------------ | Branch (2308:12): [True: 17.8k, False: 6.82k] ------------------ 2309| 17.8k| if (*ptr1 != *ptr2) { ------------------ | Branch (2309:13): [True: 32, False: 17.7k] ------------------ 2310| 32| return false; 2311| 32| } 2312| 17.7k| ptr1++; 2313| 17.7k| ptr2++; 2314| 17.7k| } 2315| | 2316| 6.82k| return true; 2317| 6.85k|} bitset_extract_setbits_avx2: 722| 992| uint32_t base) { 723| 992| if (outcapacity == 0) return 0; ------------------ | Branch (723:9): [True: 0, False: 992] ------------------ 724| 992| uint32_t *initout = out; 725| 992| __m256i baseVec = _mm256_set1_epi32(base - 1); 726| 992| __m256i incVec = _mm256_set1_epi32(64); 727| 992| __m256i add8 = _mm256_set1_epi32(8); 728| 992| uint32_t *safeout = out + outcapacity; 729| 992| size_t i = 0; 730| 829k| for (; (i < length) && (out + 64 <= safeout); ++i) { ------------------ | Branch (730:12): [True: 829k, False: 269] | Branch (730:28): [True: 828k, False: 723] ------------------ 731| 828k| uint64_t w = words[i]; 732| 828k| if (w == 0) { ------------------ | Branch (732:13): [True: 114k, False: 714k] ------------------ 733| 114k| baseVec = _mm256_add_epi32(baseVec, incVec); 734| 714k| } else { 735| 3.57M| for (int k = 0; k < 4; ++k) { ------------------ | Branch (735:29): [True: 2.85M, False: 714k] ------------------ 736| 2.85M| uint8_t byteA = (uint8_t)w; 737| 2.85M| uint8_t byteB = (uint8_t)(w >> 8); 738| 2.85M| w >>= 16; 739| 2.85M| __m256i vecA = 740| 2.85M| _mm256_loadu_si256((const __m256i *)vecDecodeTable[byteA]); 741| 2.85M| __m256i vecB = 742| 2.85M| _mm256_loadu_si256((const __m256i *)vecDecodeTable[byteB]); 743| 2.85M| uint8_t advanceA = lengthTable[byteA]; 744| 2.85M| uint8_t advanceB = lengthTable[byteB]; 745| 2.85M| vecA = _mm256_add_epi32(baseVec, vecA); 746| 2.85M| baseVec = _mm256_add_epi32(baseVec, add8); 747| 2.85M| vecB = _mm256_add_epi32(baseVec, vecB); 748| 2.85M| baseVec = _mm256_add_epi32(baseVec, add8); 749| 2.85M| _mm256_storeu_si256((__m256i *)out, vecA); 750| 2.85M| out += advanceA; 751| 2.85M| _mm256_storeu_si256((__m256i *)out, vecB); 752| 2.85M| out += advanceB; 753| 2.85M| } 754| 714k| } 755| 828k| } 756| 992| base += i * 64; 757| 15.5k| for (; (i < length) && (out < safeout); ++i) { ------------------ | Branch (757:12): [True: 15.1k, False: 369] | Branch (757:28): [True: 14.5k, False: 623] ------------------ 758| 14.5k| uint64_t w = words[i]; 759| 33.9k| while ((w != 0) && (out < safeout)) { ------------------ | Branch (759:16): [True: 19.3k, False: 14.5k] | Branch (759:28): [True: 19.3k, False: 0] ------------------ 760| 19.3k| int r = 761| 19.3k| roaring_trailing_zeroes(w); // on x64, should compile to TZCNT 762| 19.3k| uint32_t val = r + base; 763| 19.3k| memcpy(out, &val, 764| 19.3k| sizeof(uint32_t)); // should be compiled as a MOV on x64 765| 19.3k| out++; 766| 19.3k| w &= (w - 1); 767| 19.3k| } 768| 14.5k| base += 64; 769| 14.5k| } 770| 992| return out - initout; 771| 992|} bitset_extract_setbits: 776| 412| uint32_t *out, uint32_t base) { 777| 412| int outpos = 0; 778| 422k| for (size_t i = 0; i < length; ++i) { ------------------ | Branch (778:24): [True: 421k, False: 412] ------------------ 779| 421k| uint64_t w = words[i]; 780| 2.37M| while (w != 0) { ------------------ | Branch (780:16): [True: 1.95M, False: 421k] ------------------ 781| 1.95M| int r = 782| 1.95M| roaring_trailing_zeroes(w); // on x64, should compile to TZCNT 783| 1.95M| uint32_t val = r + base; 784| 1.95M| memcpy(out + outpos, &val, 785| 1.95M| sizeof(uint32_t)); // should be compiled as a MOV on x64 786| 1.95M| outpos++; 787| 1.95M| w &= (w - 1); 788| 1.95M| } 789| 421k| base += 64; 790| 421k| } 791| 412| return outpos; 792| 412|} bitset_extract_setbits_uint16: 887| 2.77k| uint16_t *out, uint16_t base) { 888| 2.77k| int outpos = 0; 889| 2.84M| for (size_t i = 0; i < length; ++i) { ------------------ | Branch (889:24): [True: 2.84M, False: 2.77k] ------------------ 890| 2.84M| uint64_t w = words[i]; 891| 5.48M| while (w != 0) { ------------------ | Branch (891:16): [True: 2.63M, False: 2.84M] ------------------ 892| 2.63M| int r = roaring_trailing_zeroes(w); 893| 2.63M| out[outpos++] = (uint16_t)(r + base); 894| 2.63M| w &= (w - 1); 895| 2.63M| } 896| 2.84M| base += 64; 897| 2.84M| } 898| 2.77k| return outpos; 899| 2.77k|} bitset_clear_list: 1067| 1.57k| uint64_t length) { 1068| 1.57k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (1068:9): [True: 1.57k, False: 0] ------------------ 1069| 1.57k| return _asm_bitset_clear_list(words, card, list, length); 1070| 1.57k| } else { 1071| 0| return _scalar_bitset_clear_list(words, card, list, length); 1072| 0| } 1073| 1.57k|} bitset_set_list_withcard: 1076| 1.30k| const uint16_t *list, uint64_t length) { 1077| 1.30k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (1077:9): [True: 1.30k, False: 0] ------------------ 1078| 1.30k| return _asm_bitset_set_list_withcard(words, card, list, length); 1079| 1.30k| } else { 1080| 0| return _scalar_bitset_set_list_withcard(words, card, list, length); 1081| 0| } 1082| 1.30k|} bitset_set_list: 1084| 193|void bitset_set_list(uint64_t *words, const uint16_t *list, uint64_t length) { 1085| 193| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (1085:9): [True: 193, False: 0] ------------------ 1086| 193| _asm_bitset_set_list(words, list, length); 1087| 193| } else { 1088| 0| _scalar_bitset_set_list(words, list, length); 1089| 0| } 1090| 193|} bitset_flip_list_withcard: 1146| 1.40k| const uint16_t *list, uint64_t length) { 1147| 1.40k| uint64_t offset, load, newload, pos, index; 1148| 1.40k| const uint16_t *end = list + length; 1149| 2.48M| while (list != end) { ------------------ | Branch (1149:12): [True: 2.48M, False: 1.40k] ------------------ 1150| 2.48M| pos = *list; 1151| 2.48M| offset = pos >> 6; 1152| 2.48M| index = pos % 64; 1153| 2.48M| load = words[offset]; 1154| 2.48M| newload = load ^ (UINT64_C(1) << index); 1155| | // todo: is a branch here all that bad? 1156| 2.48M| card += 1157| | (1 - 2 * (((UINT64_C(1) << index) & load) >> index)); // +1 or -1 1158| 2.48M| words[offset] = newload; 1159| 2.48M| list++; 1160| 2.48M| } 1161| 1.40k| return card; 1162| 1.40k|} bitset_util.c:_asm_bitset_clear_list: 988| 1.57k| uint64_t length) { 989| 1.57k| uint64_t offset, load, pos; 990| 1.57k| uint64_t shift = 6; 991| 1.57k| const uint16_t *end = list + length; 992| 1.57k| if (!length) return card; ------------------ | Branch (992:9): [True: 0, False: 1.57k] ------------------ 993| | // btr is not available as an intrinsic in GCC 994| 1.57k| __asm volatile( 995| 1.57k| "1:\n" 996| 1.57k| "movzwq (%[list]), %[pos]\n" 997| 1.57k| "shrx %[shift], %[pos], %[offset]\n" 998| 1.57k| "mov (%[words],%[offset],8), %[load]\n" 999| 1.57k| "btr %[pos], %[load]\n" 1000| 1.57k| "mov %[load], (%[words],%[offset],8)\n" 1001| 1.57k| "sbb $0, %[card]\n" 1002| 1.57k| "add $2, %[list]\n" 1003| 1.57k| "cmp %[list], %[end]\n" 1004| 1.57k| "jnz 1b" 1005| 1.57k| : [card] "+&r"(card), [list] "+&r"(list), [load] "=&r"(load), 1006| 1.57k| [pos] "=&r"(pos), [offset] "=&r"(offset) 1007| 1.57k| : [end] "r"(end), [words] "r"(words), [shift] "r"(shift) 1008| 1.57k| : 1009| 1.57k| /* clobbers */ "memory"); 1010| 1.57k| return card; 1011| 1.57k|} bitset_util.c:_asm_bitset_set_list_withcard: 906| 1.30k| uint64_t length) { 907| 1.30k| uint64_t offset, load, pos; 908| 1.30k| uint64_t shift = 6; 909| 1.30k| const uint16_t *end = list + length; 910| 1.30k| if (!length) return card; ------------------ | Branch (910:9): [True: 0, False: 1.30k] ------------------ 911| | // TODO: could unroll for performance, see bitset_set_list 912| | // bts is not available as an intrinsic in GCC 913| 1.30k| __asm volatile( 914| 1.30k| "1:\n" 915| 1.30k| "movzwq (%[list]), %[pos]\n" 916| 1.30k| "shrx %[shift], %[pos], %[offset]\n" 917| 1.30k| "mov (%[words],%[offset],8), %[load]\n" 918| 1.30k| "bts %[pos], %[load]\n" 919| 1.30k| "mov %[load], (%[words],%[offset],8)\n" 920| 1.30k| "sbb $-1, %[card]\n" 921| 1.30k| "add $2, %[list]\n" 922| 1.30k| "cmp %[list], %[end]\n" 923| 1.30k| "jnz 1b" 924| 1.30k| : [card] "+&r"(card), [list] "+&r"(list), [load] "=&r"(load), 925| 1.30k| [pos] "=&r"(pos), [offset] "=&r"(offset) 926| 1.30k| : [end] "r"(end), [words] "r"(words), [shift] "r"(shift)); 927| 1.30k| return card; 928| 1.30k|} bitset_util.c:_asm_bitset_set_list: 931| 193| uint64_t length) { 932| 193| uint64_t pos; 933| 193| const uint16_t *end = list + length; 934| | 935| 193| uint64_t shift = 6; 936| 193| uint64_t offset; 937| 193| uint64_t load; 938| 5.06k| for (; list + 3 < end; list += 4) { ------------------ | Branch (938:12): [True: 4.86k, False: 193] ------------------ 939| 4.86k| pos = list[0]; 940| 4.86k| __asm volatile( 941| 4.86k| "shrx %[shift], %[pos], %[offset]\n" 942| 4.86k| "mov (%[words],%[offset],8), %[load]\n" 943| 4.86k| "bts %[pos], %[load]\n" 944| 4.86k| "mov %[load], (%[words],%[offset],8)" 945| 4.86k| : [load] "=&r"(load), [offset] "=&r"(offset) 946| 4.86k| : [words] "r"(words), [shift] "r"(shift), [pos] "r"(pos)); 947| 4.86k| pos = list[1]; 948| 4.86k| __asm volatile( 949| 4.86k| "shrx %[shift], %[pos], %[offset]\n" 950| 4.86k| "mov (%[words],%[offset],8), %[load]\n" 951| 4.86k| "bts %[pos], %[load]\n" 952| 4.86k| "mov %[load], (%[words],%[offset],8)" 953| 4.86k| : [load] "=&r"(load), [offset] "=&r"(offset) 954| 4.86k| : [words] "r"(words), [shift] "r"(shift), [pos] "r"(pos)); 955| 4.86k| pos = list[2]; 956| 4.86k| __asm volatile( 957| 4.86k| "shrx %[shift], %[pos], %[offset]\n" 958| 4.86k| "mov (%[words],%[offset],8), %[load]\n" 959| 4.86k| "bts %[pos], %[load]\n" 960| 4.86k| "mov %[load], (%[words],%[offset],8)" 961| 4.86k| : [load] "=&r"(load), [offset] "=&r"(offset) 962| 4.86k| : [words] "r"(words), [shift] "r"(shift), [pos] "r"(pos)); 963| 4.86k| pos = list[3]; 964| 4.86k| __asm volatile( 965| 4.86k| "shrx %[shift], %[pos], %[offset]\n" 966| 4.86k| "mov (%[words],%[offset],8), %[load]\n" 967| 4.86k| "bts %[pos], %[load]\n" 968| 4.86k| "mov %[load], (%[words],%[offset],8)" 969| 4.86k| : [load] "=&r"(load), [offset] "=&r"(offset) 970| 4.86k| : [words] "r"(words), [shift] "r"(shift), [pos] "r"(pos)); 971| 4.86k| } 972| | 973| 483| while (list != end) { ------------------ | Branch (973:12): [True: 290, False: 193] ------------------ 974| 290| pos = list[0]; 975| 290| __asm volatile( 976| 290| "shrx %[shift], %[pos], %[offset]\n" 977| 290| "mov (%[words],%[offset],8), %[load]\n" 978| 290| "bts %[pos], %[load]\n" 979| 290| "mov %[load], (%[words],%[offset],8)" 980| 290| : [load] "=&r"(load), [offset] "=&r"(offset) 981| 290| : [words] "r"(words), [shift] "r"(shift), [pos] "r"(pos)); 982| 290| list++; 983| 290| } 984| 193|} array_container_create_given_capacity: 49| 141k|array_container_t *array_container_create_given_capacity(int32_t size) { 50| 141k| array_container_t *container; 51| | 52| 141k| if ((container = (array_container_t *)roaring_malloc( ------------------ | Branch (52:9): [True: 0, False: 141k] ------------------ 53| 141k| sizeof(array_container_t))) == NULL) { 54| 0| return NULL; 55| 0| } 56| | 57| 141k| if (size <= 0) { // we don't want to rely on malloc(0) ------------------ | Branch (57:9): [True: 87.0k, False: 54.1k] ------------------ 58| 87.0k| container->array = NULL; 59| 87.0k| } else if ((container->array = (uint16_t *)roaring_malloc(sizeof(uint16_t) * ------------------ | Branch (59:16): [True: 0, False: 54.1k] ------------------ 60| 54.1k| size)) == NULL) { 61| 0| roaring_free(container); 62| 0| return NULL; 63| 0| } 64| | 65| 141k| container->capacity = size; 66| 141k| container->cardinality = 0; 67| | 68| 141k| return container; 69| 141k|} array_container_create: 72| 25.3k|array_container_t *array_container_create(void) { 73| 25.3k| return array_container_create_given_capacity(ARRAY_DEFAULT_INIT_SIZE); 74| 25.3k|} array_container_create_range: 77| 126|array_container_t *array_container_create_range(uint32_t min, uint32_t max) { 78| 126| array_container_t *answer = 79| 126| array_container_create_given_capacity(max - min + 1); 80| 126| if (answer == NULL) return answer; ------------------ | Branch (80:9): [True: 0, False: 126] ------------------ 81| 126| answer->cardinality = 0; 82| 252| for (uint32_t k = min; k < max; k++) { ------------------ | Branch (82:28): [True: 126, False: 126] ------------------ 83| 126| answer->array[answer->cardinality++] = k; 84| 126| } 85| 126| return answer; 86| 126|} array_container_clone: 90| 21.6k|array_container_t *array_container_clone(const array_container_t *src) { 91| 21.6k| array_container_t *newcontainer = 92| 21.6k| array_container_create_given_capacity(src->capacity); 93| 21.6k| if (newcontainer == NULL) return NULL; ------------------ | Branch (93:9): [True: 0, False: 21.6k] ------------------ 94| | 95| 21.6k| newcontainer->cardinality = src->cardinality; 96| | 97| 21.6k| memcpy(newcontainer->array, src->array, 98| 21.6k| src->cardinality * sizeof(uint16_t)); 99| | 100| 21.6k| return newcontainer; 101| 21.6k|} array_container_shrink_to_fit: 131| 3.70k|int array_container_shrink_to_fit(array_container_t *src) { 132| 3.70k| if (src->cardinality == src->capacity) return 0; // nothing to do ------------------ | Branch (132:9): [True: 466, False: 3.23k] ------------------ 133| 3.23k| int savings = src->capacity - src->cardinality; 134| 3.23k| src->capacity = src->cardinality; 135| 3.23k| if (src->capacity == ------------------ | Branch (135:9): [True: 0, False: 3.23k] ------------------ 136| 3.23k| 0) { // we do not want to rely on realloc for zero allocs 137| 0| roaring_free(src->array); 138| 0| src->array = NULL; 139| 3.23k| } else { 140| 3.23k| uint16_t *oldarray = src->array; 141| 3.23k| src->array = (uint16_t *)roaring_realloc( 142| 3.23k| oldarray, src->capacity * sizeof(uint16_t)); 143| 3.23k| if (src->array == NULL) roaring_free(oldarray); // should never happen? ------------------ | Branch (143:13): [True: 0, False: 3.23k] ------------------ 144| 3.23k| } 145| 3.23k| return savings; 146| 3.70k|} array_container_free: 149| 141k|void array_container_free(array_container_t *arr) { 150| 141k| if (arr == NULL) return; ------------------ | Branch (150:9): [True: 0, False: 141k] ------------------ 151| 141k| roaring_free(arr->array); 152| 141k| roaring_free(arr); 153| 141k|} array_container_grow: 167| 116k| bool preserve) { 168| 116k| int32_t max = (min <= DEFAULT_MAX_SIZE ? DEFAULT_MAX_SIZE : 65536); ------------------ | Branch (168:20): [True: 116k, False: 0] ------------------ 169| 116k| int32_t new_capacity = clamp(grow_capacity(container->capacity), min, max); 170| | 171| 116k| container->capacity = new_capacity; 172| 116k| uint16_t *array = container->array; 173| | 174| 116k| if (preserve) { ------------------ | Branch (174:9): [True: 107k, False: 9.74k] ------------------ 175| 107k| container->array = 176| 107k| (uint16_t *)roaring_realloc(array, new_capacity * sizeof(uint16_t)); 177| 107k| if (container->array == NULL) roaring_free(array); ------------------ | Branch (177:13): [True: 0, False: 107k] ------------------ 178| 107k| } else { 179| 9.74k| roaring_free(array); 180| 9.74k| container->array = 181| 9.74k| (uint16_t *)roaring_malloc(new_capacity * sizeof(uint16_t)); 182| 9.74k| } 183| | 184| | // if realloc fails, we have container->array == NULL. 185| 116k|} array_container_copy: 189| 24| array_container_t *dst) { 190| 24| const int32_t cardinality = src->cardinality; 191| 24| if (cardinality > dst->capacity) { ------------------ | Branch (191:9): [True: 24, False: 0] ------------------ 192| 24| array_container_grow(dst, cardinality, false); 193| 24| } 194| | 195| 24| dst->cardinality = cardinality; 196| 24| memcpy(dst->array, src->array, cardinality * sizeof(uint16_t)); 197| 24|} array_container_union: 211| 4.59k| array_container_t *out) { 212| 4.59k| const int32_t card_1 = array_1->cardinality, card_2 = array_2->cardinality; 213| 4.59k| const int32_t max_cardinality = card_1 + card_2; 214| | 215| 4.59k| if (out->capacity < max_cardinality) { ------------------ | Branch (215:9): [True: 0, False: 4.59k] ------------------ 216| | array_container_grow(out, max_cardinality, false); 217| 0| } 218| 4.59k| out->cardinality = (int32_t)fast_union_uint16( 219| 4.59k| array_1->array, card_1, array_2->array, card_2, out->array); 220| 4.59k|} array_container_andnot: 228| 5.74k| array_container_t *out) { 229| 5.74k| if (out->capacity < array_1->cardinality) ------------------ | Branch (229:9): [True: 2.30k, False: 3.44k] ------------------ 230| 2.30k| array_container_grow(out, array_1->cardinality, false); 231| 5.74k|#if CROARING_IS_X64 232| 5.74k| if ((croaring_hardware_support() & ROARING_SUPPORTS_AVX2) && ------------------ | Branch (232:9): [True: 5.74k, False: 0] ------------------ 233| 5.74k| (out != array_1) && (out != array_2)) { ------------------ | Branch (233:9): [True: 2.30k, False: 3.44k] | Branch (233:29): [True: 2.30k, False: 0] ------------------ 234| 2.30k| out->cardinality = difference_vector16( 235| 2.30k| array_1->array, array_1->cardinality, array_2->array, 236| 2.30k| array_2->cardinality, out->array); 237| 3.44k| } else { 238| 3.44k| out->cardinality = 239| 3.44k| difference_uint16(array_1->array, array_1->cardinality, 240| 3.44k| array_2->array, array_2->cardinality, out->array); 241| 3.44k| } 242| |#else 243| | out->cardinality = 244| | difference_uint16(array_1->array, array_1->cardinality, array_2->array, 245| | array_2->cardinality, out->array); 246| |#endif 247| 5.74k|} array_container_xor: 256| 7.71k| array_container_t *out) { 257| 7.71k| const int32_t card_1 = array_1->cardinality, card_2 = array_2->cardinality; 258| 7.71k| const int32_t max_cardinality = card_1 + card_2; 259| 7.71k| if (out->capacity < max_cardinality) { ------------------ | Branch (259:9): [True: 0, False: 7.71k] ------------------ 260| 0| array_container_grow(out, max_cardinality, false); 261| 0| } 262| | 263| 7.71k|#if CROARING_IS_X64 264| 7.71k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (264:9): [True: 7.71k, False: 0] ------------------ 265| 7.71k| out->cardinality = 266| 7.71k| xor_vector16(array_1->array, array_1->cardinality, array_2->array, 267| 7.71k| array_2->cardinality, out->array); 268| 7.71k| } else { 269| 0| out->cardinality = 270| 0| xor_uint16(array_1->array, array_1->cardinality, array_2->array, 271| 0| array_2->cardinality, out->array); 272| 0| } 273| |#else 274| | out->cardinality = 275| | xor_uint16(array_1->array, array_1->cardinality, array_2->array, 276| | array_2->cardinality, out->array); 277| |#endif 278| 7.71k|} array_container_intersection: 290| 2.30k| array_container_t *out) { 291| 2.30k| int32_t card_1 = array1->cardinality, card_2 = array2->cardinality, 292| 2.30k| min_card = minimum_int32(card_1, card_2); 293| 2.30k| const int threshold = 64; // subject to tuning 294| 2.30k|#if CROARING_IS_X64 295| 2.30k| if (out->capacity < min_card) { ------------------ | Branch (295:9): [True: 2.30k, False: 0] ------------------ 296| 2.30k| array_container_grow(out, min_card + sizeof(__m128i) / sizeof(uint16_t), 297| 2.30k| false); 298| 2.30k| } 299| |#else 300| | if (out->capacity < min_card) { 301| | array_container_grow(out, min_card, false); 302| | } 303| |#endif 304| | 305| 2.30k| if (card_1 * threshold < card_2) { ------------------ | Branch (305:9): [True: 339, False: 1.96k] ------------------ 306| 339| out->cardinality = intersect_skewed_uint16( 307| 339| array1->array, card_1, array2->array, card_2, out->array); 308| 1.96k| } else if (card_2 * threshold < card_1) { ------------------ | Branch (308:16): [True: 199, False: 1.76k] ------------------ 309| 199| out->cardinality = intersect_skewed_uint16( 310| 199| array2->array, card_2, array1->array, card_1, out->array); 311| 1.76k| } else { 312| 1.76k|#if CROARING_IS_X64 313| 1.76k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (313:13): [True: 1.76k, False: 0] ------------------ 314| 1.76k| out->cardinality = intersect_vector16( 315| 1.76k| array1->array, card_1, array2->array, card_2, out->array); 316| 1.76k| } else { 317| 0| out->cardinality = intersect_uint16( 318| 0| array1->array, card_1, array2->array, card_2, out->array); 319| 0| } 320| |#else 321| | out->cardinality = intersect_uint16(array1->array, card_1, 322| | array2->array, card_2, out->array); 323| |#endif 324| 1.76k| } 325| 2.30k|} array_container_intersection_cardinality: 330| 9.21k| const array_container_t *array2) { 331| 9.21k| int32_t card_1 = array1->cardinality, card_2 = array2->cardinality; 332| 9.21k| const int threshold = 64; // subject to tuning 333| 9.21k| if (card_1 * threshold < card_2) { ------------------ | Branch (333:9): [True: 1.35k, False: 7.86k] ------------------ 334| 1.35k| return intersect_skewed_uint16_cardinality(array1->array, card_1, 335| 1.35k| array2->array, card_2); 336| 7.86k| } else if (card_2 * threshold < card_1) { ------------------ | Branch (336:16): [True: 796, False: 7.06k] ------------------ 337| 796| return intersect_skewed_uint16_cardinality(array2->array, card_2, 338| 796| array1->array, card_1); 339| 7.06k| } else { 340| 7.06k|#if CROARING_IS_X64 341| 7.06k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (341:13): [True: 7.06k, False: 0] ------------------ 342| 7.06k| return intersect_vector16_cardinality(array1->array, card_1, 343| 7.06k| array2->array, card_2); 344| 7.06k| } else { 345| 0| return intersect_uint16_cardinality(array1->array, card_1, 346| 0| array2->array, card_2); 347| 0| } 348| |#else 349| | return intersect_uint16_cardinality(array1->array, card_1, 350| | array2->array, card_2); 351| |#endif 352| 7.06k| } 353| 9.21k|} array_container_intersect: 356| 2.30k| const array_container_t *array2) { 357| 2.30k| int32_t card_1 = array1->cardinality, card_2 = array2->cardinality; 358| 2.30k| const int threshold = 64; // subject to tuning 359| 2.30k| if (card_1 * threshold < card_2) { ------------------ | Branch (359:9): [True: 339, False: 1.96k] ------------------ 360| 339| return intersect_skewed_uint16_nonempty(array1->array, card_1, 361| 339| array2->array, card_2); 362| 1.96k| } else if (card_2 * threshold < card_1) { ------------------ | Branch (362:16): [True: 199, False: 1.76k] ------------------ 363| 199| return intersect_skewed_uint16_nonempty(array2->array, card_2, 364| 199| array1->array, card_1); 365| 1.76k| } else { 366| | // we do not bother vectorizing 367| 1.76k| return intersect_uint16_nonempty(array1->array, card_1, array2->array, 368| 1.76k| card_2); 369| 1.76k| } 370| 2.30k|} array_container_intersection_inplace: 376| 4.57k| const array_container_t *src_2) { 377| 4.57k| int32_t card_1 = src_1->cardinality, card_2 = src_2->cardinality; 378| 4.57k| const int threshold = 64; // subject to tuning 379| 4.57k| if (card_1 * threshold < card_2) { ------------------ | Branch (379:9): [True: 0, False: 4.57k] ------------------ 380| 0| src_1->cardinality = intersect_skewed_uint16( 381| 0| src_1->array, card_1, src_2->array, card_2, src_1->array); 382| 4.57k| } else if (card_2 * threshold < card_1) { ------------------ | Branch (382:16): [True: 205, False: 4.36k] ------------------ 383| 205| src_1->cardinality = intersect_skewed_uint16( 384| 205| src_2->array, card_2, src_1->array, card_1, src_1->array); 385| 4.36k| } else { 386| 4.36k|#if CROARING_IS_X64 387| 4.36k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (387:13): [True: 4.36k, False: 0] ------------------ 388| 4.36k| src_1->cardinality = intersect_vector16_inplace( 389| 4.36k| src_1->array, card_1, src_2->array, card_2); 390| 4.36k| } else { 391| 0| src_1->cardinality = intersect_uint16( 392| 0| src_1->array, card_1, src_2->array, card_2, src_1->array); 393| 0| } 394| |#else 395| | src_1->cardinality = intersect_uint16( 396| | src_1->array, card_1, src_2->array, card_2, src_1->array); 397| |#endif 398| 4.36k| } 399| 4.57k|} array_container_to_uint32_array: 403| 5.62k| uint32_t base) { 404| 5.62k|#if CROARING_IS_X64 405| 5.62k| int support = croaring_hardware_support(); 406| 5.62k|#if CROARING_COMPILER_SUPPORTS_AVX512 407| 5.62k| if (support & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (407:9): [True: 0, False: 5.62k] ------------------ 408| 0| return avx512_array_container_to_uint32_array(vout, cont->array, 409| 0| cont->cardinality, base); 410| 0| } 411| 5.62k|#endif 412| 5.62k| if (support & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (412:9): [True: 5.62k, False: 0] ------------------ 413| 5.62k| return array_container_to_uint32_array_vector16( 414| 5.62k| vout, cont->array, cont->cardinality, base); 415| 5.62k| } 416| 0|#endif // CROARING_IS_X64 417| 0| int outpos = 0; 418| 0| uint32_t *out = (uint32_t *)vout; 419| 0| size_t i = 0; 420| 0| for (; i < (size_t)cont->cardinality; ++i) { ------------------ | Branch (420:12): [True: 0, False: 0] ------------------ 421| 0| const uint32_t val = base + cont->array[i]; 422| 0| memcpy(out + outpos, &val, 423| 0| sizeof(uint32_t)); // should be compiled as a MOV on x64 424| 0| outpos++; 425| 0| } 426| 0| return outpos; 427| 5.62k|} array_container_number_of_runs: 495| 12.4k|int32_t array_container_number_of_runs(const array_container_t *ac) { 496| | // Can SIMD work here? 497| 12.4k| int32_t nr_runs = 0; 498| 12.4k| int32_t prev = -2; 499| 1.61M| for (const uint16_t *p = ac->array; p != ac->array + ac->cardinality; ++p) { ------------------ | Branch (499:41): [True: 1.60M, False: 12.4k] ------------------ 500| 1.60M| if (*p != prev + 1) nr_runs++; ------------------ | Branch (500:13): [True: 867k, False: 733k] ------------------ 501| 1.60M| prev = *p; 502| 1.60M| } 503| 12.4k| return nr_runs; 504| 12.4k|} array_container_write: 512| 6.82k|int32_t array_container_write(const array_container_t *container, char *buf) { 513| |#if CROARING_IS_BIG_ENDIAN 514| | for (int32_t i = 0; i < container->cardinality; ++i) { 515| | uint16_t v_le = croaring_htole16(container->array[i]); 516| | memcpy(buf + i * sizeof(uint16_t), &v_le, sizeof(uint16_t)); 517| | } 518| |#else 519| 6.82k| memcpy(buf, container->array, container->cardinality * sizeof(uint16_t)); 520| 6.82k|#endif 521| 6.82k| return array_container_size_in_bytes(container); 522| 6.82k|} array_container_is_subset: 525| 5.59k| const array_container_t *container2) { 526| 5.59k| if (container1->cardinality > container2->cardinality) { ------------------ | Branch (526:9): [True: 638, False: 4.96k] ------------------ 527| 638| return false; 528| 638| } 529| 4.96k| int i1 = 0, i2 = 0; 530| 49.3k| while (i1 < container1->cardinality && i2 < container2->cardinality) { ------------------ | Branch (530:12): [True: 49.3k, False: 0] | Branch (530:44): [True: 49.2k, False: 24] ------------------ 531| 49.2k| if (container1->array[i1] == container2->array[i2]) { ------------------ | Branch (531:13): [True: 0, False: 49.2k] ------------------ 532| 0| i1++; 533| 0| i2++; 534| 49.2k| } else if (container1->array[i1] > container2->array[i2]) { ------------------ | Branch (534:20): [True: 44.3k, False: 4.93k] ------------------ 535| 44.3k| i2++; 536| 44.3k| } else { // container1->array[i1] < container2->array[i2] 537| 4.93k| return false; 538| 4.93k| } 539| 49.2k| } 540| 24| if (i1 == container1->cardinality) { ------------------ | Branch (540:9): [True: 0, False: 24] ------------------ 541| 0| return true; 542| 24| } else { 543| | return false; 544| 24| } 545| 24|} array_container_read: 548| 6.86k| const char *buf) { 549| 6.86k| if (container->capacity < cardinality) { ------------------ | Branch (549:9): [True: 0, False: 6.86k] ------------------ 550| 0| array_container_grow(container, cardinality, false); 551| 0| } 552| 6.86k| container->cardinality = cardinality; 553| |#if CROARING_IS_BIG_ENDIAN 554| | for (int32_t i = 0; i < cardinality; ++i) { 555| | uint16_t v_le; 556| | memcpy(&v_le, buf + i * sizeof(uint16_t), sizeof(uint16_t)); 557| | container->array[i] = croaring_letoh16(v_le); 558| | } 559| |#else 560| 6.86k| memcpy(container->array, buf, container->cardinality * sizeof(uint16_t)); 561| 6.86k|#endif 562| | 563| 6.86k| return array_container_size_in_bytes(container); 564| 6.86k|} array_container_iterate: 567| 6.22k| roaring_iterator iterator, void *ptr) { 568| 2.00M| for (int i = 0; i < cont->cardinality; i++) ------------------ | Branch (568:21): [True: 2.00M, False: 6.22k] ------------------ 569| 2.00M| if (!iterator(cont->array[i] + base, ptr)) return false; ------------------ | Branch (569:13): [True: 0, False: 2.00M] ------------------ 570| 6.22k| return true; 571| 6.22k|} array.c:clamp: 162| 116k|static inline int32_t clamp(int32_t val, int32_t min, int32_t max) { 163| 116k| return ((val < min) ? min : (val > max) ? max : val); ------------------ | Branch (163:13): [True: 25.9k, False: 91.0k] | Branch (163:33): [True: 241, False: 90.8k] ------------------ 164| 116k|} array.c:grow_capacity: 155| 116k|static inline int32_t grow_capacity(int32_t capacity) { 156| 116k| return (capacity <= 0) ? ARRAY_DEFAULT_INIT_SIZE ------------------ | Branch (156:12): [True: 25.3k, False: 91.6k] ------------------ 157| 116k| : capacity < 64 ? capacity * 2 ------------------ | Branch (157:14): [True: 66.5k, False: 25.0k] ------------------ 158| 91.6k| : capacity < 1024 ? capacity * 3 / 2 ------------------ | Branch (158:14): [True: 24.7k, False: 278] ------------------ 159| 25.0k| : capacity * 5 / 4; 160| 116k|} array.c:minimum_int32: 280| 2.30k|static inline int32_t minimum_int32(int32_t a, int32_t b) { 281| 2.30k| return (a < b) ? a : b; ------------------ | Branch (281:12): [True: 958, False: 1.34k] ------------------ 282| 2.30k|} bitset_container_clear: 51| 72.1k|void bitset_container_clear(bitset_container_t *bitset) { 52| 72.1k| memset(bitset->words, 0, sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 53| 72.1k| bitset->cardinality = 0; 54| 72.1k|} bitset_container_set_all: 56| 155|void bitset_container_set_all(bitset_container_t *bitset) { 57| | memset(bitset->words, INT64_C(-1), 58| 155| sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 59| 155| bitset->cardinality = (1 << 16); 60| 155|} bitset_container_create: 91| 72.1k|bitset_container_t *bitset_container_create(void) { 92| 72.1k| bitset_container_t *bitset = bitset_container_allocate(); 93| 72.1k| if (bitset) { ------------------ | Branch (93:9): [True: 72.1k, False: 0] ------------------ 94| 72.1k| bitset_container_clear(bitset); 95| 72.1k| } 96| 72.1k| return bitset; 97| 72.1k|} bitset_container_create_uninitialized: 99| 66.9k|bitset_container_t *bitset_container_create_uninitialized(void) { 100| 66.9k| bitset_container_t *bitset = bitset_container_allocate(); 101| 66.9k| if (bitset) { ------------------ | Branch (101:9): [True: 66.9k, False: 0] ------------------ 102| 66.9k| bitset->cardinality = 0; 103| 66.9k| } 104| 66.9k| return bitset; 105| 66.9k|} bitset_container_copy: 109| 3.65k| bitset_container_t *dest) { 110| 3.65k| dest->cardinality = source->cardinality; 111| 3.65k| memcpy(dest->words, source->words, 112| 3.65k| sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 113| 3.65k|} bitset_container_free: 144| 273k|void bitset_container_free(bitset_container_t *bitset) { 145| 273k| if (bitset == NULL) return; ------------------ | Branch (145:9): [True: 0, False: 273k] ------------------ 146| 273k| roaring_aligned_free(bitset->words); 147| 273k| roaring_free(bitset); 148| 273k|} bitset_container_clone: 152| 134k|bitset_container_t *bitset_container_clone(const bitset_container_t *src) { 153| 134k| bitset_container_t *bitset = 154| 134k| (bitset_container_t *)roaring_malloc(sizeof(bitset_container_t)); 155| | 156| 134k| if (!bitset) { ------------------ | Branch (156:9): [True: 0, False: 134k] ------------------ 157| 0| return NULL; 158| 0| } 159| | 160| 134k| size_t align_size = 32; 161| 134k|#if CROARING_IS_X64 162| 134k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (162:9): [True: 0, False: 134k] ------------------ 163| | // sizeof(__m512i) == 64 164| 0| align_size = 64; 165| 134k| } else { 166| | // sizeof(__m256i) == 32 167| 134k| align_size = 32; 168| 134k| } 169| 134k|#endif 170| 134k| bitset->words = (uint64_t *)roaring_aligned_malloc( 171| 134k| align_size, sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 172| 134k| if (!bitset->words) { ------------------ | Branch (172:9): [True: 0, False: 134k] ------------------ 173| 0| roaring_free(bitset); 174| 0| return NULL; 175| 0| } 176| 134k| bitset->cardinality = src->cardinality; 177| 134k| memcpy(bitset->words, src->words, 178| 134k| sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 179| 134k| return bitset; 180| 134k|} bitset_container_compute_cardinality: 284| 3.22k|int bitset_container_compute_cardinality(const bitset_container_t *bitset) { 285| 3.22k| int support = croaring_hardware_support(); 286| 3.22k|#if CROARING_COMPILER_SUPPORTS_AVX512 287| 3.22k| if (support & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (287:9): [True: 0, False: 3.22k] ------------------ 288| 0| return (int)avx512_vpopcount( 289| 0| (const __m512i *)bitset->words, 290| 0| BITSET_CONTAINER_SIZE_IN_WORDS / (WORDS_IN_AVX512_REG)); ------------------ | | 268| 0|#define WORDS_IN_AVX512_REG sizeof(__m512i) / sizeof(uint64_t) ------------------ 291| 0| } else 292| 3.22k|#endif // CROARING_COMPILER_SUPPORTS_AVX512 293| 3.22k| if (support & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (293:13): [True: 3.22k, False: 0] ------------------ 294| 3.22k| return (int)avx2_harley_seal_popcount256( 295| 3.22k| (const __m256i *)bitset->words, 296| 3.22k| BITSET_CONTAINER_SIZE_IN_WORDS / (CROARING_WORDS_IN_AVX2_REG)); ------------------ | | 265| 3.22k|#define CROARING_WORDS_IN_AVX2_REG sizeof(__m256i) / sizeof(uint64_t) ------------------ 297| 3.22k| } else { 298| 0| return _scalar_bitset_container_compute_cardinality(bitset); 299| 0| } 300| 3.22k|} bitset_container_and_nocard: 739| 1.40k| bitset_container_t *dst) { \ 740| 1.40k| int support = croaring_hardware_support(); \ 741| 1.40k| if ( support & ROARING_SUPPORTS_AVX512 ) { \ ------------------ | Branch (741:10): [True: 0, False: 1.40k] ------------------ 742| 0| return _avx512_bitset_container_##opname##_nocard(src_1, src_2, dst); \ 743| 0| } \ 744| 1.40k| else if ( support & ROARING_SUPPORTS_AVX2 ) { \ ------------------ | Branch (744:15): [True: 1.40k, False: 0] ------------------ 745| 1.40k| return _avx2_bitset_container_##opname##_nocard(src_1, src_2, dst); \ 746| 1.40k| } else { \ 747| 0| return _scalar_bitset_container_##opname##_nocard(src_1, src_2, dst); \ 748| 0| } \ 749| 1.40k| } \ bitset_container_and_justcard: 751| 1.40k| const bitset_container_t *src_2) { \ 752| 1.40k| int support = croaring_hardware_support(); \ 753| 1.40k| if ( support & ROARING_SUPPORTS_AVX512 ) { \ ------------------ | Branch (753:10): [True: 0, False: 1.40k] ------------------ 754| 0| return _avx512_bitset_container_##opname##_justcard(src_1, src_2); \ 755| 0| } \ 756| 1.40k| else if ( support & ROARING_SUPPORTS_AVX2 ) { \ ------------------ | Branch (756:15): [True: 1.40k, False: 0] ------------------ 757| 1.40k| return _avx2_bitset_container_##opname##_justcard(src_1, src_2); \ 758| 1.40k| } else { \ 759| 0| return _scalar_bitset_container_##opname##_justcard(src_1, src_2); \ 760| 0| } \ 761| 1.40k| } bitset_container_xor: 726| 1.45k| bitset_container_t *dst) { \ 727| 1.45k| int support = croaring_hardware_support(); \ 728| 1.45k| if ( support & ROARING_SUPPORTS_AVX512 ) { \ ------------------ | Branch (728:10): [True: 0, False: 1.45k] ------------------ 729| 0| return _avx512_bitset_container_##opname(src_1, src_2, dst); \ 730| 0| } \ 731| 1.45k| else if ( support & ROARING_SUPPORTS_AVX2 ) { \ ------------------ | Branch (731:15): [True: 1.45k, False: 0] ------------------ 732| 1.45k| return _avx2_bitset_container_##opname(src_1, src_2, dst); \ 733| 1.45k| } else { \ 734| 0| return _scalar_bitset_container_##opname(src_1, src_2, dst); \ 735| 0| } \ 736| 1.45k| } \ bitset_container_to_uint32_array: 951| 1.40k|){ 952| 1.40k|#if CROARING_IS_X64 953| 1.40k| int support = croaring_hardware_support(); 954| 1.40k|#if CROARING_COMPILER_SUPPORTS_AVX512 955| | // Unlike the AVX2 kernel, the AVX-512 one needs no cardinality heuristic: 956| | // it skips empty words and only stores the blocks that carry a value, so it 957| | // beat the scalar loop at every cardinality measured from 16 to 65536 (on 958| | // an Emerald Rapids Xeon, 1.3x at 512 values, 6x at 4096, 7.5x at 8192). 959| 1.40k| if( support & ROARING_SUPPORTS_AVX512 ) ------------------ | Branch (959:8): [True: 0, False: 1.40k] ------------------ 960| 0| return (int) bitset_extract_setbits_avx512(bc->words, 961| 0| BITSET_CONTAINER_SIZE_IN_WORDS, out, bc->cardinality, base); 962| 1.40k| else 963| 1.40k|#endif 964| 1.40k| if(( support & ROARING_SUPPORTS_AVX2 ) && (bc->cardinality >= 8192)) // heuristic ------------------ | Branch (964:7): [True: 1.40k, False: 0] | Branch (964:47): [True: 992, False: 412] ------------------ 965| 992| return (int) bitset_extract_setbits_avx2(bc->words, 966| 992| BITSET_CONTAINER_SIZE_IN_WORDS, out, bc->cardinality, base); 967| 412| else 968| 412| return (int) bitset_extract_setbits(bc->words, 969| 412| BITSET_CONTAINER_SIZE_IN_WORDS, out, base); 970| |#else 971| | return (int) bitset_extract_setbits(bc->words, 972| | BITSET_CONTAINER_SIZE_IN_WORDS, out, base); 973| |#endif 974| 1.40k|} bitset_container_write: 1069| 66.9k| char *buf) { 1070| |#if CROARING_IS_BIG_ENDIAN 1071| | for (int32_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; ++i) { 1072| | uint64_t w_le = croaring_htole64(container->words[i]); 1073| | memcpy(buf + i * sizeof(uint64_t), &w_le, sizeof(uint64_t)); 1074| | } 1075| |#else 1076| 66.9k| memcpy(buf, container->words, BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t)); 1077| 66.9k|#endif 1078| 66.9k| return bitset_container_size_in_bytes(container); 1079| 66.9k|} bitset_container_read: 1083| 66.9k| const char *buf) { 1084| 66.9k| container->cardinality = cardinality; 1085| |#if CROARING_IS_BIG_ENDIAN 1086| | for (int32_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; ++i) { 1087| | uint64_t w_le; 1088| | memcpy(&w_le, buf + i * sizeof(uint64_t), sizeof(uint64_t)); 1089| | container->words[i] = croaring_letoh64(w_le); 1090| | } 1091| |#else 1092| 66.9k| memcpy(container->words, buf, BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t)); 1093| 66.9k|#endif 1094| 66.9k| return bitset_container_size_in_bytes(container); 1095| 66.9k|} bitset_container_iterate: 1097| 1.56k|bool bitset_container_iterate(const bitset_container_t *cont, uint32_t base, roaring_iterator iterator, void *ptr) { 1098| 1.60M| for (int32_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; ++i ) { ------------------ | Branch (1098:23): [True: 1.60M, False: 1.56k] ------------------ 1099| 1.60M| uint64_t w = cont->words[i]; 1100| 61.4M| while (w != 0) { ------------------ | Branch (1100:12): [True: 59.8M, False: 1.60M] ------------------ 1101| 59.8M| uint64_t t = w & (~w + 1); 1102| 59.8M| int r = roaring_trailing_zeroes(w); 1103| 59.8M| if(!iterator(r + base, ptr)) return false; ------------------ | Branch (1103:10): [True: 0, False: 59.8M] ------------------ 1104| 59.8M| w ^= t; 1105| 59.8M| } 1106| 1.60M| base += 64; 1107| 1.60M| } 1108| 1.56k| return true; 1109| 1.56k|} bitset_container_equals: 1163| 66.9k|bool bitset_container_equals(const bitset_container_t *container1, const bitset_container_t *container2) { 1164| 66.9k| if((container1->cardinality != BITSET_UNKNOWN_CARDINALITY) && (container2->cardinality != BITSET_UNKNOWN_CARDINALITY)) { ------------------ | Branch (1164:6): [True: 66.9k, False: 0] | Branch (1164:65): [True: 66.9k, False: 0] ------------------ 1165| 66.9k| if(container1->cardinality != container2->cardinality) { ------------------ | Branch (1165:8): [True: 0, False: 66.9k] ------------------ 1166| 0| return false; 1167| 0| } 1168| 66.9k| if (container1->cardinality == INT32_C(0x10000)) { ------------------ | Branch (1168:9): [True: 63.2k, False: 3.76k] ------------------ 1169| 63.2k| return true; 1170| 63.2k| } 1171| 66.9k| } 1172| 3.76k|#if CROARING_IS_X64 1173| 3.76k| int support = croaring_hardware_support(); 1174| 3.76k|#if CROARING_COMPILER_SUPPORTS_AVX512 1175| 3.76k| if( support & ROARING_SUPPORTS_AVX512 ) { ------------------ | Branch (1175:7): [True: 0, False: 3.76k] ------------------ 1176| 0| return _avx512_bitset_container_equals(container1, container2); 1177| 0| } 1178| 3.76k| else 1179| 3.76k|#endif 1180| 3.76k| if( support & ROARING_SUPPORTS_AVX2 ) { ------------------ | Branch (1180:7): [True: 3.76k, False: 0] ------------------ 1181| 3.76k| return _avx2_bitset_container_equals(container1, container2); 1182| 3.76k| } 1183| 0|#endif 1184| 0| return memcmp(container1->words, 1185| 0| container2->words, 1186| 0| BITSET_CONTAINER_SIZE_IN_WORDS*sizeof(uint64_t)) == 0; 1187| 3.76k|} bitset_container_minimum: 1237| 1.15k|uint16_t bitset_container_minimum(const bitset_container_t *container) { 1238| 10.1k| for (int32_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; ++i ) { ------------------ | Branch (1238:23): [True: 10.1k, False: 0] ------------------ 1239| 10.1k| uint64_t w = container->words[i]; 1240| 10.1k| if (w != 0) { ------------------ | Branch (1240:9): [True: 1.15k, False: 9.01k] ------------------ 1241| 1.15k| int r = roaring_trailing_zeroes(w); 1242| 1.15k| return r + i * 64; 1243| 1.15k| } 1244| 10.1k| } 1245| 0| return UINT16_MAX; 1246| 1.15k|} bitset_container_maximum: 1249| 4.23k|uint16_t bitset_container_maximum(const bitset_container_t *container) { 1250| 990k| for (int32_t i = BITSET_CONTAINER_SIZE_IN_WORDS - 1; i > 0; --i ) { ------------------ | Branch (1250:56): [True: 990k, False: 0] ------------------ 1251| 990k| uint64_t w = container->words[i]; 1252| 990k| if (w != 0) { ------------------ | Branch (1252:9): [True: 4.23k, False: 986k] ------------------ 1253| 4.23k| int r = roaring_leading_zeroes(w); 1254| 4.23k| return i * 64 + 63 - r; 1255| 4.23k| } 1256| 990k| } 1257| 0| return 0; 1258| 4.23k|} bitset_container_index_equalorlarger: 1318| 1.79k|int bitset_container_index_equalorlarger(const bitset_container_t *container, uint16_t x) { 1319| 1.79k| uint32_t x32 = x; 1320| 1.79k| uint32_t k = x32 / 64; 1321| 1.79k| uint64_t word = container->words[k]; 1322| 1.79k| const int diff = x32 - k * 64; // in [0,64) 1323| 1.79k| word = (word >> diff) << diff; // a mask is faster, but we don't care 1324| 4.82k| while(word == 0) { ------------------ | Branch (1324:9): [True: 3.03k, False: 1.79k] ------------------ 1325| 3.03k| k++; 1326| 3.03k| if(k == BITSET_CONTAINER_SIZE_IN_WORDS) return -1; ------------------ | Branch (1326:8): [True: 0, False: 3.03k] ------------------ 1327| 3.03k| word = container->words[k]; 1328| 3.03k| } 1329| 1.79k| return k * 64 + roaring_trailing_zeroes(word); 1330| 1.79k|} bitset.c:bitset_container_allocate: 62| 139k|static bitset_container_t *bitset_container_allocate(void) { 63| 139k| bitset_container_t *bitset = 64| 139k| (bitset_container_t *)roaring_malloc(sizeof(bitset_container_t)); 65| | 66| 139k| if (!bitset) { ------------------ | Branch (66:9): [True: 0, False: 139k] ------------------ 67| 0| return NULL; 68| 0| } 69| | 70| 139k| size_t align_size = 32; 71| 139k|#if CROARING_IS_X64 72| 139k| int support = croaring_hardware_support(); 73| 139k| if (support & ROARING_SUPPORTS_AVX512) { ------------------ | Branch (73:9): [True: 0, False: 139k] ------------------ 74| | // sizeof(__m512i) == 64 75| 0| align_size = 64; 76| 139k| } else { 77| | // sizeof(__m256i) == 32 78| 139k| align_size = 32; 79| 139k| } 80| 139k|#endif 81| 139k| bitset->words = (uint64_t *)roaring_aligned_malloc( 82| 139k| align_size, sizeof(uint64_t) * BITSET_CONTAINER_SIZE_IN_WORDS); 83| 139k| if (!bitset->words) { ------------------ | Branch (83:9): [True: 0, False: 139k] ------------------ 84| 0| roaring_free(bitset); 85| 0| return NULL; 86| 0| } 87| 139k| return bitset; 88| 139k|} bitset.c:_avx2_bitset_container_and_nocard: 518| 1.40k| bitset_container_t *dst) { \ 519| 1.40k| const uint8_t *__restrict__ words_1 = (const uint8_t *)src_1->words; \ 520| 1.40k| const uint8_t *__restrict__ words_2 = (const uint8_t *)src_2->words; \ 521| 1.40k| /* not using the blocking optimization for some reason*/ \ 522| 1.40k| uint8_t *out = (uint8_t *)dst->words; \ 523| 1.40k| const int innerloop = 8; \ 524| 1.40k| for (size_t i = 0; \ 525| 46.3k| i < BITSET_CONTAINER_SIZE_IN_WORDS / (CROARING_WORDS_IN_AVX2_REG); \ ------------------ | | 265| 46.3k|#define CROARING_WORDS_IN_AVX2_REG sizeof(__m256i) / sizeof(uint64_t) ------------------ | Branch (525:10): [True: 44.9k, False: 1.40k] ------------------ 526| 44.9k| i += innerloop) { \ 527| 44.9k| __m256i A1, A2, AO; \ 528| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1)); \ 529| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2)); \ 530| 44.9k| AO = avx_intrinsic(A2, A1); \ 531| 44.9k| _mm256_storeu_si256((__m256i *)out, AO); \ 532| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 32)); \ 533| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 32)); \ 534| 44.9k| AO = avx_intrinsic(A2, A1); \ 535| 44.9k| _mm256_storeu_si256((__m256i *)(out + 32), AO); \ 536| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 64)); \ 537| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 64)); \ 538| 44.9k| AO = avx_intrinsic(A2, A1); \ 539| 44.9k| _mm256_storeu_si256((__m256i *)(out + 64), AO); \ 540| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 96)); \ 541| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 96)); \ 542| 44.9k| AO = avx_intrinsic(A2, A1); \ 543| 44.9k| _mm256_storeu_si256((__m256i *)(out + 96), AO); \ 544| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 128)); \ 545| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 128)); \ 546| 44.9k| AO = avx_intrinsic(A2, A1); \ 547| 44.9k| _mm256_storeu_si256((__m256i *)(out + 128), AO); \ 548| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 160)); \ 549| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 160)); \ 550| 44.9k| AO = avx_intrinsic(A2, A1); \ 551| 44.9k| _mm256_storeu_si256((__m256i *)(out + 160), AO); \ 552| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 192)); \ 553| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 192)); \ 554| 44.9k| AO = avx_intrinsic(A2, A1); \ 555| 44.9k| _mm256_storeu_si256((__m256i *)(out + 192), AO); \ 556| 44.9k| A1 = _mm256_lddqu_si256((const __m256i *)(words_1 + 224)); \ 557| 44.9k| A2 = _mm256_lddqu_si256((const __m256i *)(words_2 + 224)); \ 558| 44.9k| AO = avx_intrinsic(A2, A1); \ 559| 44.9k| _mm256_storeu_si256((__m256i *)(out + 224), AO); \ 560| 44.9k| out += 256; \ 561| 44.9k| words_1 += 256; \ 562| 44.9k| words_2 += 256; \ 563| 44.9k| } \ 564| 1.40k| dst->cardinality = BITSET_UNKNOWN_CARDINALITY; \ 565| 1.40k| return dst->cardinality; \ 566| 1.40k| } bitset.c:_avx2_bitset_container_and_justcard: 587| 1.40k| const bitset_container_t *src_1, const bitset_container_t *src_2) { \ 588| 1.40k| const __m256i *__restrict__ data1 = (const __m256i *)src_1->words; \ 589| 1.40k| const __m256i *__restrict__ data2 = (const __m256i *)src_2->words; \ 590| 1.40k| return (int)avx2_harley_seal_popcount256_##opname( \ 591| 1.40k| data2, data1, BITSET_CONTAINER_SIZE_IN_WORDS / (CROARING_WORDS_IN_AVX2_REG)); \ ------------------ | | 265| 1.40k|#define CROARING_WORDS_IN_AVX2_REG sizeof(__m256i) / sizeof(uint64_t) ------------------ 592| 1.40k| } bitset.c:_avx2_bitset_container_xor: 573| 1.45k| bitset_container_t *dst) { \ 574| 1.45k| const __m256i *__restrict__ words_1 = (const __m256i *)src_1->words; \ 575| 1.45k| const __m256i *__restrict__ words_2 = (const __m256i *)src_2->words; \ 576| 1.45k| __m256i *out = (__m256i *)dst->words; \ 577| 1.45k| dst->cardinality = (int32_t)avx2_harley_seal_popcount256andstore_##opname( \ 578| 1.45k| words_2, words_1, out, \ 579| 1.45k| BITSET_CONTAINER_SIZE_IN_WORDS / (CROARING_WORDS_IN_AVX2_REG)); \ ------------------ | | 265| 1.45k|#define CROARING_WORDS_IN_AVX2_REG sizeof(__m256i) / sizeof(uint64_t) ------------------ 580| 1.45k| return dst->cardinality; \ 581| 1.45k| } \ bitset.c:_avx2_bitset_container_equals: 1146| 3.76k|static inline bool _avx2_bitset_container_equals(const bitset_container_t *container1, const bitset_container_t *container2) { 1147| 3.76k| const __m256i *ptr1 = (const __m256i*)container1->words; 1148| 3.76k| const __m256i *ptr2 = (const __m256i*)container2->words; 1149| 967k| for (size_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS*sizeof(uint64_t)/32; i++) { ------------------ | Branch (1149:24): [True: 963k, False: 3.76k] ------------------ 1150| 963k| __m256i r1 = _mm256_loadu_si256(ptr1+i); 1151| 963k| __m256i r2 = _mm256_loadu_si256(ptr2+i); 1152| 963k| int mask = _mm256_movemask_epi8(_mm256_cmpeq_epi8(r1, r2)); 1153| 963k| if ((uint32_t)mask != UINT32_MAX) { ------------------ | Branch (1153:11): [True: 0, False: 963k] ------------------ 1154| 0| return false; 1155| 0| } 1156| 963k| } 1157| 3.76k| return true; 1158| 3.76k|} container_free: 58| 584k|void container_free(container_t *c, uint8_t type) { 59| 584k| switch (type) { 60| 271k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 271k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (60:9): [True: 271k, False: 313k] ------------------ 61| 271k| bitset_container_free(CAST_bitset(c)); ------------------ | | 52| 271k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 271k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 62| 271k| break; 63| 127k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 127k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (63:9): [True: 127k, False: 457k] ------------------ 64| 127k| array_container_free(CAST_array(c)); ------------------ | | 54| 127k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 127k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 65| 127k| break; 66| 186k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 186k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (66:9): [True: 186k, False: 398k] ------------------ 67| 186k| run_container_free(CAST_run(c)); ------------------ | | 77| 186k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 186k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 68| 186k| break; 69| 0| case SHARED_CONTAINER_TYPE: ------------------ | | 51| 0|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (69:9): [True: 0, False: 584k] ------------------ 70| 0| shared_container_free(CAST_shared(c)); ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 71| 0| break; 72| 0| default: ------------------ | Branch (72:9): [True: 0, False: 584k] ------------------ 73| 0| assert(false); 74| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 75| 584k| } 76| 584k|} get_copy_of_container: 181| 236| bool copy_on_write) { 182| 236| if (copy_on_write) { ------------------ | Branch (182:9): [True: 0, False: 236] ------------------ 183| 0| shared_container_t *shared_container; 184| 0| if (*typecode == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 0|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (184:13): [True: 0, False: 0] ------------------ 185| 0| shared_container = CAST_shared(c); ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 186| 0| croaring_refcount_inc(&shared_container->counter); 187| 0| return shared_container; 188| 0| } 189| 0| assert(*typecode != SHARED_CONTAINER_TYPE); 190| | 191| 0| if ((shared_container = (shared_container_t *)roaring_malloc( ------------------ | Branch (191:13): [True: 0, False: 0] ------------------ 192| 0| sizeof(shared_container_t))) == NULL) { 193| 0| return NULL; 194| 0| } 195| | 196| 0| shared_container->container = c; 197| 0| shared_container->typecode = *typecode; 198| | // At this point, we are creating new shared container 199| | // so there should be no other references, and setting 200| | // the counter to 2 - even non-atomically - is safe as 201| | // long as the value is set before the return statement. 202| 0| shared_container->counter = 2; 203| 0| *typecode = SHARED_CONTAINER_TYPE; ------------------ | | 51| 0|#define SHARED_CONTAINER_TYPE 4 ------------------ 204| | 205| 0| return shared_container; 206| 0| } // copy_on_write 207| | // otherwise, no copy on write... 208| 236| const container_t *actual_container = container_unwrap_shared(c, typecode); 209| | assert(*typecode != SHARED_CONTAINER_TYPE); 210| 236| return container_clone(actual_container, *typecode); 211| 236|} container_clone: 217| 321k|container_t *container_clone(const container_t *c, uint8_t typecode) { 218| | // We do not want to allow cloning of shared containers. 219| | // c = container_unwrap_shared(c, &typecode); 220| 321k| switch (typecode) { 221| 134k| case BITSET_CONTAINER_TYPE: ------------------ | | 48| 134k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (221:9): [True: 134k, False: 186k] ------------------ 222| 134k| return bitset_container_clone(const_CAST_bitset(c)); ------------------ | | 53| 134k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 134k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 223| 21.6k| case ARRAY_CONTAINER_TYPE: ------------------ | | 49| 21.6k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (223:9): [True: 21.6k, False: 299k] ------------------ 224| 21.6k| return array_container_clone(const_CAST_array(c)); ------------------ | | 55| 21.6k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 21.6k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 225| 164k| case RUN_CONTAINER_TYPE: ------------------ | | 50| 164k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (225:9): [True: 164k, False: 156k] ------------------ 226| 164k| return run_container_clone(const_CAST_run(c)); ------------------ | | 78| 164k|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 164k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 227| 0| case SHARED_CONTAINER_TYPE: ------------------ | | 51| 0|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (227:9): [True: 0, False: 321k] ------------------ 228| | // Shared containers are not cloneable. Are you mixing COW and 229| | // non-COW bitmaps? 230| 0| return NULL; 231| 0| default: ------------------ | Branch (231:9): [True: 0, False: 321k] ------------------ 232| 0| assert(false); 233| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 234| 0| return NULL; 235| 321k| } 236| 321k|} container_init_iterator: 305| 18.5k| uint16_t *value) { 306| 18.5k| switch (typecode) { 307| 3.62k| case BITSET_CONTAINER_TYPE: { ------------------ | | 48| 3.62k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (307:9): [True: 3.62k, False: 14.8k] ------------------ 308| 3.62k| const bitset_container_t *bc = const_CAST_bitset(c); ------------------ | | 53| 3.62k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 3.62k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 309| 3.62k| uint32_t wordindex = 0; 310| 3.62k| uint64_t word; 311| 24.3k| while ((word = bc->words[wordindex]) == 0) { ------------------ | Branch (311:20): [True: 20.7k, False: 3.62k] ------------------ 312| 20.7k| wordindex++; 313| 20.7k| } 314| | // word is non-zero 315| 3.62k| int32_t index = wordindex * 64 + roaring_trailing_zeroes(word); 316| 3.62k| *value = index; 317| 3.62k| return ROARING_INIT_ROARING_CONTAINER_ITERATOR_T{ ------------------ | | 13| 3.62k|#define ROARING_INIT_ROARING_CONTAINER_ITERATOR_T (roaring_container_iterator_t) ------------------ 318| 3.62k| .index = index, 319| 3.62k| }; 320| 0| } 321| 14.3k| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 14.3k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (321:9): [True: 14.3k, False: 4.11k] ------------------ 322| 14.3k| const array_container_t *ac = const_CAST_array(c); ------------------ | | 55| 14.3k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 14.3k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 323| 14.3k| *value = ac->array[0]; 324| 14.3k| return ROARING_INIT_ROARING_CONTAINER_ITERATOR_T{ ------------------ | | 13| 14.3k|#define ROARING_INIT_ROARING_CONTAINER_ITERATOR_T (roaring_container_iterator_t) ------------------ 325| 14.3k| .index = 0, 326| 14.3k| }; 327| 0| } 328| 489| case RUN_CONTAINER_TYPE: { ------------------ | | 50| 489|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (328:9): [True: 489, False: 18.0k] ------------------ 329| 489| const run_container_t *rc = const_CAST_run(c); ------------------ | | 78| 489|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 489|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 330| 489| *value = rc->runs[0].value; 331| 489| return ROARING_INIT_ROARING_CONTAINER_ITERATOR_T{ ------------------ | | 13| 489|#define ROARING_INIT_ROARING_CONTAINER_ITERATOR_T (roaring_container_iterator_t) ------------------ 332| 489| .index = 0, 333| 489| }; 334| 0| } 335| 0| default: ------------------ | Branch (335:9): [True: 0, False: 18.5k] ------------------ 336| 0| assert(false); 337| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 338| 0| return ROARING_INIT_ROARING_CONTAINER_ITERATOR_T{0}; ------------------ | | 13| 0|#define ROARING_INIT_ROARING_CONTAINER_ITERATOR_T (roaring_container_iterator_t) ------------------ 339| 18.5k| } 340| 18.5k|} container_iterator_lower_bound: 388| 6.17k| uint16_t *value_out, uint16_t val) { 389| 6.17k| if (val > container_maximum(c, typecode)) { ------------------ | Branch (389:9): [True: 0, False: 6.17k] ------------------ 390| 0| return false; 391| 0| } 392| 6.17k| switch (typecode) { 393| 1.79k| case BITSET_CONTAINER_TYPE: { ------------------ | | 48| 1.79k|#define BITSET_CONTAINER_TYPE 1 ------------------ | Branch (393:9): [True: 1.79k, False: 4.37k] ------------------ 394| 1.79k| const bitset_container_t *bc = const_CAST_bitset(c); ------------------ | | 53| 1.79k|#define const_CAST_bitset(c) CAST(const bitset_container_t *, c) | | ------------------ | | | | 86| 1.79k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 395| 1.79k| it->index = bitset_container_index_equalorlarger(bc, val); 396| 1.79k| *value_out = it->index; 397| 1.79k| return true; 398| 0| } 399| 4.37k| case ARRAY_CONTAINER_TYPE: { ------------------ | | 49| 4.37k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (399:9): [True: 4.37k, False: 1.79k] ------------------ 400| 4.37k| const array_container_t *ac = const_CAST_array(c); ------------------ | | 55| 4.37k|#define const_CAST_array(c) CAST(const array_container_t *, c) | | ------------------ | | | | 86| 4.37k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 401| 4.37k| it->index = array_container_index_equalorlarger(ac, val); 402| 4.37k| *value_out = ac->array[it->index]; 403| 4.37k| return true; 404| 0| } 405| 0| case RUN_CONTAINER_TYPE: { ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (405:9): [True: 0, False: 6.17k] ------------------ 406| 0| const run_container_t *rc = const_CAST_run(c); ------------------ | | 78| 0|#define const_CAST_run(c) CAST(const run_container_t *, c) | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 407| 0| it->index = run_container_index_equalorlarger(rc, val); 408| 0| if (rc->runs[it->index].value <= val) { ------------------ | Branch (408:17): [True: 0, False: 0] ------------------ 409| 0| *value_out = val; 410| 0| } else { 411| 0| *value_out = rc->runs[it->index].value; 412| 0| } 413| 0| return true; 414| 0| } 415| 0| default: ------------------ | Branch (415:9): [True: 0, False: 6.17k] ------------------ 416| 0| assert(false); 417| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 418| 0| return false; 419| 6.17k| } 420| 6.17k|} bitset_container_from_array: 22| 2.21k|bitset_container_t *bitset_container_from_array(const array_container_t *ac) { 23| 2.21k| bitset_container_t *ans = bitset_container_create(); 24| 2.21k| int limit = array_container_cardinality(ac); 25| 2.28M| for (int i = 0; i < limit; ++i) bitset_container_set(ans, ac->array[i]); ------------------ | Branch (25:21): [True: 2.28M, False: 2.21k] ------------------ 26| 2.21k| return ans; 27| 2.21k|} bitset_container_from_run: 29| 574|bitset_container_t *bitset_container_from_run(const run_container_t *arr) { 30| 574| int card = run_container_cardinality(arr); 31| 574| bitset_container_t *answer = bitset_container_create(); 32| 22.1k| for (int rlepos = 0; rlepos < arr->n_runs; ++rlepos) { ------------------ | Branch (32:26): [True: 21.5k, False: 574] ------------------ 33| 21.5k| rle16_t vl = arr->runs[rlepos]; 34| 21.5k| bitset_set_lenrange(answer->words, vl.value, vl.length); 35| 21.5k| } 36| 574| answer->cardinality = card; 37| 574| return answer; 38| 574|} array_container_from_run: 40| 1.15k|array_container_t *array_container_from_run(const run_container_t *arr) { 41| 1.15k| array_container_t *answer = 42| 1.15k| array_container_create_given_capacity(run_container_cardinality(arr)); 43| 1.15k| answer->cardinality = 0; 44| 83.8k| for (int rlepos = 0; rlepos < arr->n_runs; ++rlepos) { ------------------ | Branch (44:26): [True: 82.6k, False: 1.15k] ------------------ 45| 82.6k| int run_start = arr->runs[rlepos].value; 46| 82.6k| int run_end = run_start + arr->runs[rlepos].length; 47| | 48| 441k| for (int run_value = run_start; run_value <= run_end; ++run_value) { ------------------ | Branch (48:41): [True: 358k, False: 82.6k] ------------------ 49| 358k| answer->array[answer->cardinality++] = (uint16_t)run_value; 50| 358k| } 51| 82.6k| } 52| 1.15k| return answer; 53| 1.15k|} array_container_from_bitset: 55| 2.70k|array_container_t *array_container_from_bitset(const bitset_container_t *bits) { 56| 2.70k| array_container_t *result = 57| 2.70k| array_container_create_given_capacity(bits->cardinality); 58| 2.70k| result->cardinality = bits->cardinality; 59| 2.70k|#if CROARING_IS_X64 60| 2.70k|#if CROARING_COMPILER_SUPPORTS_AVX512 61| | // Every caller first checks that the cardinality fits an array container 62| | // (<= DEFAULT_MAX_SIZE), so the input here is sparse: under 6.25% density, 63| | // a couple of set bits per word. 64| | // Below ~1024 values the vector decoder does not make back its per-word 65| | // cost and the scalar loop is as fast or faster; from 1024 up it wins 66| | // (1.2x at 1024, 1.4x at 4096, measured on an Emerald Rapids Xeon). 67| 2.70k| if ((bits->cardinality >= 1024) && ------------------ | Branch (67:9): [True: 571, False: 2.13k] ------------------ 68| 571| (croaring_hardware_support() & ROARING_SUPPORTS_AVX512)) { ------------------ | Branch (68:9): [True: 0, False: 571] ------------------ 69| 0| bitset_extract_setbits_avx512_uint16( 70| 0| bits->words, BITSET_CONTAINER_SIZE_IN_WORDS, result->array, 71| 0| bits->cardinality, 0); 72| 0| } else 73| 2.70k|#endif 74| 2.70k| { 75| | // sse version ends up being slower here 76| | // (bitset_extract_setbits_sse_uint16) 77| | // because of the sparsity of the data 78| 2.70k| bitset_extract_setbits_uint16( 79| 2.70k| bits->words, BITSET_CONTAINER_SIZE_IN_WORDS, result->array, 0); 80| 2.70k| } 81| |#else 82| | // If the system is not x64, then we have no accelerated function. 83| | bitset_extract_setbits_uint16(bits->words, BITSET_CONTAINER_SIZE_IN_WORDS, 84| | result->array, 0); 85| |#endif 86| | 87| 2.70k| return result; 88| 2.70k|} convert_to_bitset_or_array_container: 127| 66.5k| uint8_t *resulttype) { 128| 66.5k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (128:9): [True: 1.19k, False: 65.3k] ------------------ 129| 1.19k| array_container_t *answer = array_container_create_given_capacity(card); 130| 1.19k| answer->cardinality = 0; 131| 63.5k| for (int rlepos = 0; rlepos < rc->n_runs; ++rlepos) { ------------------ | Branch (131:30): [True: 62.3k, False: 1.19k] ------------------ 132| 62.3k| uint16_t run_start = rc->runs[rlepos].value; 133| 62.3k| uint16_t run_end = run_start + rc->runs[rlepos].length; 134| 620k| for (uint16_t run_value = run_start; run_value < run_end; ------------------ | Branch (134:50): [True: 557k, False: 62.3k] ------------------ 135| 557k| ++run_value) { 136| 557k| answer->array[answer->cardinality++] = run_value; 137| 557k| } 138| 62.3k| answer->array[answer->cardinality++] = run_end; 139| 62.3k| } 140| 1.19k| assert(card == answer->cardinality); 141| 1.19k| *resulttype = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.19k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 142| | // run_container_free(r); 143| 1.19k| return answer; 144| 1.19k| } 145| 65.3k| bitset_container_t *answer = bitset_container_create(); 146| 150k| for (int rlepos = 0; rlepos < rc->n_runs; ++rlepos) { ------------------ | Branch (146:26): [True: 85.4k, False: 65.3k] ------------------ 147| 85.4k| uint16_t run_start = rc->runs[rlepos].value; 148| 85.4k| bitset_set_lenrange(answer->words, run_start, rc->runs[rlepos].length); 149| 85.4k| } 150| 65.3k| answer->cardinality = card; 151| 65.3k| *resulttype = BITSET_CONTAINER_TYPE; ------------------ | | 48| 65.3k|#define BITSET_CONTAINER_TYPE 1 ------------------ 152| | // run_container_free(r); 153| 65.3k| return answer; 154| 66.5k|} convert_run_to_efficient_container: 162| 130k| uint8_t *typecode_after) { 163| 130k| int32_t size_as_run_container = 164| 130k| run_container_serialized_size_in_bytes(c->n_runs); 165| | 166| 130k| int32_t size_as_bitset_container = 167| 130k| bitset_container_serialized_size_in_bytes(); 168| 130k| int32_t card = run_container_cardinality(c); 169| 130k| int32_t size_as_array_container = 170| 130k| array_container_serialized_size_in_bytes(card); 171| | 172| 130k| int32_t min_size_non_run = 173| 130k| size_as_bitset_container < size_as_array_container ------------------ | Branch (173:9): [True: 58.0k, False: 72.3k] ------------------ 174| 130k| ? size_as_bitset_container 175| 130k| : size_as_array_container; 176| 130k| if (size_as_run_container <= min_size_non_run) { // no conversion ------------------ | Branch (176:9): [True: 66.3k, False: 64.0k] ------------------ 177| 66.3k| *typecode_after = RUN_CONTAINER_TYPE; ------------------ | | 50| 66.3k|#define RUN_CONTAINER_TYPE 3 ------------------ 178| 66.3k| return c; 179| 66.3k| } 180| 64.0k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (180:9): [True: 64.0k, False: 0] ------------------ 181| | // to array 182| 64.0k| array_container_t *answer = array_container_create_given_capacity(card); 183| 64.0k| answer->cardinality = 0; 184| 209k| for (int rlepos = 0; rlepos < c->n_runs; ++rlepos) { ------------------ | Branch (184:30): [True: 145k, False: 64.0k] ------------------ 185| 145k| int run_start = c->runs[rlepos].value; 186| 145k| int run_end = run_start + c->runs[rlepos].length; 187| | 188| 389k| for (int run_value = run_start; run_value <= run_end; ++run_value) { ------------------ | Branch (188:45): [True: 243k, False: 145k] ------------------ 189| 243k| answer->array[answer->cardinality++] = (uint16_t)run_value; 190| 243k| } 191| 145k| } 192| 64.0k| *typecode_after = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 64.0k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 193| 64.0k| return answer; 194| 64.0k| } 195| | 196| | // else to bitset 197| 0| bitset_container_t *answer = bitset_container_create(); 198| | 199| 0| for (int rlepos = 0; rlepos < c->n_runs; ++rlepos) { ------------------ | Branch (199:26): [True: 0, False: 0] ------------------ 200| 0| int start = c->runs[rlepos].value; 201| 0| int end = start + c->runs[rlepos].length; 202| 0| bitset_set_range(answer->words, start, end + 1); 203| 0| } 204| 0| answer->cardinality = card; 205| 0| *typecode_after = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 206| 0| return answer; 207| 64.0k|} convert_run_to_efficient_container_and_free: 211| 117k| run_container_t *c, uint8_t *typecode_after) { 212| 117k| container_t *answer = convert_run_to_efficient_container(c, typecode_after); 213| 117k| if (answer != c) run_container_free(c); ------------------ | Branch (213:9): [True: 56.8k, False: 60.9k] ------------------ 214| 117k| return answer; 215| 117k|} convert_run_optimize: 225| 12.4k| uint8_t *typecode_after) { 226| 12.4k| if (typecode_original == RUN_CONTAINER_TYPE) { ------------------ | | 50| 12.4k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (226:9): [True: 0, False: 12.4k] ------------------ 227| 0| container_t *newc = 228| 0| convert_run_to_efficient_container(CAST_run(c), typecode_after); ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 229| 0| if (newc != c) { ------------------ | Branch (229:13): [True: 0, False: 0] ------------------ 230| 0| container_free(c, typecode_original); 231| 0| } 232| 0| return newc; 233| 12.4k| } else if (typecode_original == ARRAY_CONTAINER_TYPE) { ------------------ | | 49| 12.4k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (233:16): [True: 12.4k, False: 0] ------------------ 234| | // it might need to be converted to a run container. 235| 12.4k| array_container_t *c_qua_array = CAST_array(c); ------------------ | | 54| 12.4k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 12.4k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 236| 12.4k| int32_t n_runs = array_container_number_of_runs(c_qua_array); 237| 12.4k| int32_t size_as_run_container = 238| 12.4k| run_container_serialized_size_in_bytes(n_runs); 239| 12.4k| int32_t card = array_container_cardinality(c_qua_array); 240| 12.4k| int32_t size_as_array_container = 241| 12.4k| array_container_serialized_size_in_bytes(card); 242| | 243| 12.4k| if (size_as_run_container >= size_as_array_container) { ------------------ | Branch (243:13): [True: 8.51k, False: 3.97k] ------------------ 244| 8.51k| *typecode_after = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 8.51k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 245| 8.51k| return c; 246| 8.51k| } 247| | // else convert array to run container 248| 3.97k| run_container_t *answer = run_container_create_given_capacity(n_runs); 249| 3.97k| int prev = -2; 250| 3.97k| int run_start = -1; 251| | 252| 3.97k| assert(card > 0); 253| 657k| for (int i = 0; i < card; ++i) { ------------------ | Branch (253:25): [True: 653k, False: 3.97k] ------------------ 254| 653k| uint16_t cur_val = c_qua_array->array[i]; 255| 653k| if (cur_val != prev + 1) { ------------------ | Branch (255:17): [True: 265k, False: 388k] ------------------ 256| | // new run starts; flush old one, if any 257| 265k| if (run_start != -1) add_run(answer, run_start, prev); ------------------ | Branch (257:21): [True: 261k, False: 3.97k] ------------------ 258| 265k| run_start = cur_val; 259| 265k| } 260| 653k| prev = c_qua_array->array[i]; 261| 653k| } 262| 3.97k| assert(run_start >= 0); 263| | // now prev is the last seen value 264| 3.97k| add_run(answer, run_start, prev); 265| 3.97k| *typecode_after = RUN_CONTAINER_TYPE; ------------------ | | 50| 3.97k|#define RUN_CONTAINER_TYPE 3 ------------------ 266| 3.97k| array_container_free(c_qua_array); 267| 3.97k| return answer; 268| 12.4k| } else if (typecode_original == ------------------ | Branch (268:16): [True: 0, False: 0] ------------------ 269| 0| BITSET_CONTAINER_TYPE) { // run conversions on bitset ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 270| | // does bitset need conversion to run? 271| 0| bitset_container_t *c_qua_bitset = CAST_bitset(c); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 272| 0| int32_t n_runs = bitset_container_number_of_runs(c_qua_bitset); 273| 0| int32_t size_as_run_container = 274| 0| run_container_serialized_size_in_bytes(n_runs); 275| 0| int32_t size_as_bitset_container = 276| 0| bitset_container_serialized_size_in_bytes(); 277| | 278| 0| if (size_as_bitset_container <= size_as_run_container) { ------------------ | Branch (278:13): [True: 0, False: 0] ------------------ 279| | // no conversion needed. 280| 0| *typecode_after = BITSET_CONTAINER_TYPE; ------------------ | | 48| 0|#define BITSET_CONTAINER_TYPE 1 ------------------ 281| 0| return c; 282| 0| } 283| | // bitset to runcontainer (ported from Java RunContainer( 284| | // BitmapContainer bc, int nbrRuns)) 285| 0| assert(n_runs > 0); // no empty bitmaps 286| 0| run_container_t *answer = run_container_create_given_capacity(n_runs); 287| | 288| 0| int long_ctr = 0; 289| 0| uint64_t cur_word = c_qua_bitset->words[0]; 290| 0| while (true) { ------------------ | Branch (290:16): [True: 0, Folded] ------------------ 291| 0| while (cur_word == UINT64_C(0) && ------------------ | Branch (291:20): [True: 0, False: 0] ------------------ 292| 0| long_ctr < BITSET_CONTAINER_SIZE_IN_WORDS - 1) ------------------ | Branch (292:20): [True: 0, False: 0] ------------------ 293| 0| cur_word = c_qua_bitset->words[++long_ctr]; 294| | 295| 0| if (cur_word == UINT64_C(0)) { ------------------ | Branch (295:17): [True: 0, False: 0] ------------------ 296| 0| bitset_container_free(c_qua_bitset); 297| 0| *typecode_after = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 298| 0| return answer; 299| 0| } 300| | 301| 0| int local_run_start = roaring_trailing_zeroes(cur_word); 302| 0| int run_start = local_run_start + 64 * long_ctr; 303| 0| uint64_t cur_word_with_1s = cur_word | (cur_word - 1); 304| | 305| 0| int run_end = 0; 306| 0| while (cur_word_with_1s == UINT64_C(0xFFFFFFFFFFFFFFFF) && ------------------ | Branch (306:20): [True: 0, False: 0] ------------------ 307| 0| long_ctr < BITSET_CONTAINER_SIZE_IN_WORDS - 1) ------------------ | Branch (307:20): [True: 0, False: 0] ------------------ 308| 0| cur_word_with_1s = c_qua_bitset->words[++long_ctr]; 309| | 310| 0| if (cur_word_with_1s == UINT64_C(0xFFFFFFFFFFFFFFFF)) { ------------------ | Branch (310:17): [True: 0, False: 0] ------------------ 311| 0| run_end = 64 + long_ctr * 64; // exclusive, I guess 312| 0| add_run(answer, run_start, run_end - 1); 313| 0| bitset_container_free(c_qua_bitset); 314| 0| *typecode_after = RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 315| 0| return answer; 316| 0| } 317| 0| int local_run_end = roaring_trailing_zeroes(~cur_word_with_1s); 318| 0| run_end = local_run_end + long_ctr * 64; 319| 0| add_run(answer, run_start, run_end - 1); 320| 0| cur_word = cur_word_with_1s & (cur_word_with_1s + 1); 321| 0| } 322| 0| return answer; 323| 0| } else { 324| 0| assert(false); 325| 0| roaring_unreachable; ------------------ | | 240| 0|#define roaring_unreachable __builtin_unreachable() ------------------ 326| 0| return NULL; 327| 0| } 328| 12.4k|} convert.c:add_run: 91| 265k|static void add_run(run_container_t *rc, int s, int e) { 92| 265k| rc->runs[rc->n_runs].value = s; 93| 265k| rc->runs[rc->n_runs].length = e - s; 94| 265k| rc->n_runs++; 95| 265k|} array_bitset_container_andnot: 26| 531| array_container_t *dst) { 27| | // follows Java implementation as of June 2016 28| 531| if (dst->capacity < src_1->cardinality) { ------------------ | Branch (28:9): [True: 531, False: 0] ------------------ 29| 531| array_container_grow(dst, src_1->cardinality, false); 30| 531| } 31| 531| int32_t newcard = 0; 32| 531| const int32_t origcard = src_1->cardinality; 33| 91.2k| for (int i = 0; i < origcard; ++i) { ------------------ | Branch (33:21): [True: 90.7k, False: 531] ------------------ 34| 90.7k| uint16_t key = src_1->array[i]; 35| 90.7k| dst->array[newcard] = key; 36| 90.7k| newcard += 1 - bitset_container_contains(src_2, key); 37| 90.7k| } 38| 531| dst->cardinality = newcard; 39| 531|} bitset_array_container_iandnot: 83| 1.42k| container_t **dst) { 84| 1.42k| *dst = src_1; 85| 1.42k| src_1->cardinality = 86| 1.42k| (int32_t)bitset_clear_list(src_1->words, (uint64_t)src_1->cardinality, 87| 1.42k| src_2->array, (uint64_t)src_2->cardinality); 88| | 89| 1.42k| if (src_1->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (89:9): [True: 190, False: 1.23k] ------------------ 90| 190| *dst = array_container_from_bitset(src_1); 91| 190| bitset_container_free(src_1); 92| 190| return false; // not bitset 93| 190| } else 94| 1.23k| return true; 95| 1.42k|} run_bitset_container_andnot: 106| 613| container_t **dst) { 107| | // follows the Java implementation as of June 2016 108| 613| int card = run_container_cardinality(src_1); 109| 613| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (109:9): [True: 613, False: 0] ------------------ 110| | // must be an array 111| 613| array_container_t *answer = array_container_create_given_capacity(card); 112| 613| answer->cardinality = 0; 113| 51.0k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (113:34): [True: 50.4k, False: 613] ------------------ 114| 50.4k| rle16_t rle = src_1->runs[rlepos]; 115| 184k| for (int run_value = rle.value; run_value <= rle.value + rle.length; ------------------ | Branch (115:45): [True: 134k, False: 50.4k] ------------------ 116| 134k| ++run_value) { 117| 134k| if (!bitset_container_get(src_2, (uint16_t)run_value)) { ------------------ | Branch (117:21): [True: 36.4k, False: 97.9k] ------------------ 118| 36.4k| answer->array[answer->cardinality++] = (uint16_t)run_value; 119| 36.4k| } 120| 134k| } 121| 50.4k| } 122| 613| *dst = answer; 123| 613| return false; 124| 613| } else { // we guess it will be a bitset, though have to check guess when 125| | // done 126| 0| bitset_container_t *answer = bitset_container_clone(src_2); 127| | 128| 0| uint32_t last_pos = 0; 129| 0| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (129:34): [True: 0, False: 0] ------------------ 130| 0| rle16_t rle = src_1->runs[rlepos]; 131| | 132| 0| uint32_t start = rle.value; 133| 0| uint32_t end = start + rle.length + 1; 134| 0| bitset_reset_range(answer->words, last_pos, start); 135| 0| bitset_flip_range(answer->words, start, end); 136| 0| last_pos = end; 137| 0| } 138| 0| bitset_reset_range(answer->words, last_pos, (uint32_t)(1 << 16)); 139| | 140| 0| answer->cardinality = bitset_container_compute_cardinality(answer); 141| | 142| 0| if (answer->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (142:13): [True: 0, False: 0] ------------------ 143| 0| *dst = array_container_from_bitset(answer); 144| 0| bitset_container_free(answer); 145| 0| return false; // not bitset 146| 0| } 147| 0| *dst = answer; 148| | return true; // bitset 149| 0| } 150| 613|} run_array_container_andnot: 279| 1.94k| container_t **dst) { 280| | // follows the Java impl as of June 2016 281| | 282| 1.94k| int card = run_container_cardinality(src_1); 283| 1.94k| const int arbitrary_threshold = 32; 284| | 285| 1.94k| if (card <= arbitrary_threshold) { ------------------ | Branch (285:9): [True: 368, False: 1.58k] ------------------ 286| 368| if (src_2->cardinality == 0) { ------------------ | Branch (286:13): [True: 0, False: 368] ------------------ 287| 0| *dst = run_container_clone(src_1); 288| 0| return RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 289| 0| } 290| | // Java's "lazyandNot.toEfficientContainer" thing 291| 368| run_container_t *answer = run_container_create_given_capacity( 292| 368| card + array_container_cardinality(src_2)); 293| | 294| 368| int rlepos = 0; 295| 368| int xrlepos = 0; // "x" is src_2 296| 368| rle16_t rle = src_1->runs[rlepos]; 297| 368| int32_t start = rle.value; 298| 368| int32_t end = start + rle.length + 1; 299| 368| int32_t xstart = src_2->array[xrlepos]; 300| | 301| 13.2k| while ((rlepos < src_1->n_runs) && (xrlepos < src_2->cardinality)) { ------------------ | Branch (301:16): [True: 12.9k, False: 307] | Branch (301:44): [True: 12.9k, False: 61] ------------------ 302| 12.9k| if (end <= xstart) { ------------------ | Branch (302:17): [True: 641, False: 12.2k] ------------------ 303| | // output the first run 304| 641| answer->runs[answer->n_runs++] = 305| 641| CROARING_MAKE_RLE16(start, end - start - 1); ------------------ | | 60| 641| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 306| 641| rlepos++; 307| 641| if (rlepos < src_1->n_runs) { ------------------ | Branch (307:21): [True: 586, False: 55] ------------------ 308| 586| start = src_1->runs[rlepos].value; 309| 586| end = start + src_1->runs[rlepos].length + 1; 310| 586| } 311| 12.2k| } else if (xstart + 1 <= start) { ------------------ | Branch (311:24): [True: 8.29k, False: 3.97k] ------------------ 312| | // exit the second run 313| 8.29k| xrlepos++; 314| 8.29k| if (xrlepos < src_2->cardinality) { ------------------ | Branch (314:21): [True: 8.23k, False: 61] ------------------ 315| 8.23k| xstart = src_2->array[xrlepos]; 316| 8.23k| } 317| 8.29k| } else { 318| 3.97k| if (start < xstart) { ------------------ | Branch (318:21): [True: 331, False: 3.64k] ------------------ 319| 331| answer->runs[answer->n_runs++] = 320| 331| CROARING_MAKE_RLE16(start, xstart - start - 1); ------------------ | | 60| 331| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 321| 331| } 322| 3.97k| if (xstart + 1 < end) { ------------------ | Branch (322:21): [True: 2.39k, False: 1.57k] ------------------ 323| 2.39k| start = xstart + 1; 324| 2.39k| } else { 325| 1.57k| rlepos++; 326| 1.57k| if (rlepos < src_1->n_runs) { ------------------ | Branch (326:25): [True: 1.32k, False: 252] ------------------ 327| 1.32k| start = src_1->runs[rlepos].value; 328| 1.32k| end = start + src_1->runs[rlepos].length + 1; 329| 1.32k| } 330| 1.57k| } 331| 3.97k| } 332| 12.9k| } 333| 368| if (rlepos < src_1->n_runs) { ------------------ | Branch (333:13): [True: 61, False: 307] ------------------ 334| 61| answer->runs[answer->n_runs++] = 335| 61| CROARING_MAKE_RLE16(start, end - start - 1); ------------------ | | 60| 61| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 336| 61| rlepos++; 337| 61| if (rlepos < src_1->n_runs) { ------------------ | Branch (337:17): [True: 19, False: 42] ------------------ 338| 19| memcpy(answer->runs + answer->n_runs, src_1->runs + rlepos, 339| 19| (src_1->n_runs - rlepos) * sizeof(rle16_t)); 340| 19| answer->n_runs += (src_1->n_runs - rlepos); 341| 19| } 342| 61| } 343| 368| uint8_t return_type; 344| 368| *dst = convert_run_to_efficient_container(answer, &return_type); 345| 368| if (answer != *dst) run_container_free(answer); ------------------ | Branch (345:13): [True: 230, False: 138] ------------------ 346| 368| return return_type; 347| 368| } 348| | // else it's a bitmap or array 349| | 350| 1.58k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (350:9): [True: 1.27k, False: 308] ------------------ 351| 1.27k| array_container_t *ac = array_container_create_given_capacity(card); 352| | // nb Java code used a generic iterator-based merge to compute 353| | // difference 354| 1.27k| ac->cardinality = run_array_array_subtract(src_1, src_2, ac); 355| 1.27k| *dst = ac; 356| 1.27k| return ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.27k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 357| 1.27k| } 358| 308| bitset_container_t *ans = bitset_container_from_run(src_1); 359| 308| bool result_is_bitset = bitset_array_container_iandnot(ans, src_2, dst); 360| 308| return (result_is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE); ------------------ | | 48| 285|#define BITSET_CONTAINER_TYPE 1 ------------------ return (result_is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE); ------------------ | | 49| 23|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (360:13): [True: 285, False: 23] ------------------ 361| 1.58k|} run_array_container_iandnot: 372| 742| container_t **dst) { 373| | // dummy implementation same as June 2016 Java 374| 742| int ans = run_array_container_andnot(src_1, src_2, dst); 375| 742| run_container_free(src_1); 376| 742| return ans; 377| 742|} array_run_container_andnot: 383| 809| array_container_t *dst) { 384| | // basically following Java impl as of June 2016 385| 809| if (src_1->cardinality > dst->capacity) { ------------------ | Branch (385:9): [True: 759, False: 50] ------------------ 386| 759| array_container_grow(dst, src_1->cardinality, false); 387| 759| } 388| | 389| 809| if (src_2->n_runs == 0) { ------------------ | Branch (389:9): [True: 0, False: 809] ------------------ 390| 0| memmove(dst->array, src_1->array, 391| 0| sizeof(uint16_t) * src_1->cardinality); 392| 0| dst->cardinality = src_1->cardinality; 393| 0| return; 394| 0| } 395| 809| int32_t run_start = src_2->runs[0].value; 396| 809| int32_t run_end = run_start + src_2->runs[0].length; 397| 809| int which_run = 0; 398| | 399| 809| uint16_t val = 0; 400| 809| int dest_card = 0; 401| 176k| for (int i = 0; i < src_1->cardinality; ++i) { ------------------ | Branch (401:21): [True: 176k, False: 809] ------------------ 402| 176k| val = src_1->array[i]; 403| 176k| if (val < run_start) ------------------ | Branch (403:13): [True: 47.3k, False: 128k] ------------------ 404| 47.3k| dst->array[dest_card++] = val; 405| 128k| else if (val <= run_end) { ------------------ | Branch (405:18): [True: 93.2k, False: 35.5k] ------------------ 406| 93.2k| ; // omitted item 407| 93.2k| } else { 408| 48.0k| do { 409| 48.0k| if (which_run + 1 < src_2->n_runs) { ------------------ | Branch (409:21): [True: 47.8k, False: 188] ------------------ 410| 47.8k| ++which_run; 411| 47.8k| run_start = src_2->runs[which_run].value; 412| 47.8k| run_end = run_start + src_2->runs[which_run].length; 413| | 414| 47.8k| } else 415| 188| run_start = run_end = (1 << 16) + 1; 416| 48.0k| } while (val > run_end); ------------------ | Branch (416:22): [True: 12.5k, False: 35.5k] ------------------ 417| 35.5k| --i; 418| 35.5k| } 419| 176k| } 420| 809| dst->cardinality = dest_card; 421| 809|} array_run_container_iandnot: 428| 50| const run_container_t *src_2) { 429| 50| array_run_container_andnot(src_1, src_2, src_1); 430| 50|} run_run_container_andnot: 437| 1.13k| const run_container_t *src_2, container_t **dst) { 438| 1.13k| run_container_t *ans = run_container_create(); 439| 1.13k| run_container_andnot(src_1, src_2, ans); 440| 1.13k| uint8_t typecode_after; 441| 1.13k| *dst = convert_run_to_efficient_container_and_free(ans, &typecode_after); 442| 1.13k| return typecode_after; 443| 1.13k|} run_run_container_iandnot: 453| 463| const run_container_t *src_2, container_t **dst) { 454| | // following Java impl as of June 2016 (dummy) 455| 463| int ans = run_run_container_andnot(src_1, src_2, dst); 456| 463| run_container_free(src_1); 457| 463| return ans; 458| 463|} array_array_container_andnot: 466| 2.30k| array_container_t *dst) { 467| 2.30k| array_container_andnot(src_1, src_2, dst); 468| 2.30k|} array_array_container_iandnot: 473| 3.44k| const array_container_t *src_2) { 474| 3.44k| array_container_andnot(src_1, src_2, src_1); 475| 3.44k|} mixed_andnot.c:run_array_array_subtract: 235| 1.27k| array_container_t *a_out) { 236| 1.27k| int out_card = 0; 237| 1.27k| int32_t in_array_pos = 238| 1.27k| -1; // since advanceUntil always assumes we start the search AFTER this 239| | 240| 95.4k| for (int rlepos = 0; rlepos < rc->n_runs; rlepos++) { ------------------ | Branch (240:26): [True: 94.2k, False: 1.27k] ------------------ 241| 94.2k| int32_t start = rc->runs[rlepos].value; 242| 94.2k| int32_t end = start + rc->runs[rlepos].length + 1; 243| | 244| 94.2k| in_array_pos = advanceUntil(a_in->array, in_array_pos, 245| 94.2k| a_in->cardinality, (uint16_t)start); 246| | 247| 94.2k| if (in_array_pos >= a_in->cardinality) { // run has no items subtracted ------------------ | Branch (247:13): [True: 5.22k, False: 88.9k] ------------------ 248| 90.2k| for (int32_t i = start; i < end; ++i) ------------------ | Branch (248:37): [True: 85.0k, False: 5.22k] ------------------ 249| 85.0k| a_out->array[out_card++] = (uint16_t)i; 250| 88.9k| } else { 251| 88.9k| uint16_t next_nonincluded = a_in->array[in_array_pos]; 252| 88.9k| if (next_nonincluded >= end) { ------------------ | Branch (252:17): [True: 29.6k, False: 59.3k] ------------------ 253| | // another case when run goes unaltered 254| 79.2k| for (int32_t i = start; i < end; ++i) ------------------ | Branch (254:41): [True: 49.6k, False: 29.6k] ------------------ 255| 49.6k| a_out->array[out_card++] = (uint16_t)i; 256| 29.6k| in_array_pos--; // ensure we see this item again if necessary 257| 59.3k| } else { 258| 283k| for (int32_t i = start; i < end; ++i) ------------------ | Branch (258:41): [True: 223k, False: 59.3k] ------------------ 259| 223k| if (i != next_nonincluded) ------------------ | Branch (259:25): [True: 93.3k, False: 130k] ------------------ 260| 93.3k| a_out->array[out_card++] = (uint16_t)i; 261| 130k| else // 0 should ensure we don't match 262| 130k| next_nonincluded = 263| 130k| (in_array_pos + 1 >= a_in->cardinality) ------------------ | Branch (263:29): [True: 1.01k, False: 129k] ------------------ 264| 130k| ? 0 265| 130k| : a_in->array[++in_array_pos]; 266| 59.3k| in_array_pos--; // see again 267| 59.3k| } 268| 88.9k| } 269| 94.2k| } 270| 1.27k| return out_card; 271| 1.27k|} array_container_equal_bitset: 10| 457| const bitset_container_t* container2) { 11| 457| if (container2->cardinality != BITSET_UNKNOWN_CARDINALITY) { ------------------ | Branch (11:9): [True: 457, False: 0] ------------------ 12| 457| if (container2->cardinality != container1->cardinality) { ------------------ | Branch (12:13): [True: 457, False: 0] ------------------ 13| 457| return false; 14| 457| } 15| 457| } 16| 0| int32_t pos = 0; 17| 0| for (int32_t i = 0; i < BITSET_CONTAINER_SIZE_IN_WORDS; ++i) { ------------------ | Branch (17:25): [True: 0, False: 0] ------------------ 18| 0| uint64_t w = container2->words[i]; 19| 0| while (w != 0) { ------------------ | Branch (19:16): [True: 0, False: 0] ------------------ 20| 0| uint64_t t = w & (~w + 1); 21| 0| uint16_t r = i * 64 + roaring_trailing_zeroes(w); 22| 0| if (pos >= container1->cardinality) { ------------------ | Branch (22:17): [True: 0, False: 0] ------------------ 23| 0| return false; 24| 0| } 25| 0| if (container1->array[pos] != r) { ------------------ | Branch (25:17): [True: 0, False: 0] ------------------ 26| 0| return false; 27| 0| } 28| 0| ++pos; 29| 0| w ^= t; 30| 0| } 31| 0| } 32| 0| return (pos == container1->cardinality); 33| 0|} run_container_equals_array: 36| 897| const array_container_t* container2) { 37| 897| if (run_container_cardinality(container1) != container2->cardinality) ------------------ | Branch (37:9): [True: 851, False: 46] ------------------ 38| 851| return false; 39| 46| int32_t pos = 0; 40| 46| for (int i = 0; i < container1->n_runs; ++i) { ------------------ | Branch (40:21): [True: 46, False: 0] ------------------ 41| 46| const uint32_t run_start = container1->runs[i].value; 42| 46| const uint32_t le = container1->runs[i].length; 43| | 44| 46| if (container2->array[pos] != run_start) { ------------------ | Branch (44:13): [True: 46, False: 0] ------------------ 45| 46| return false; 46| 46| } 47| | 48| 0| if (container2->array[pos + le] != run_start + le) { ------------------ | Branch (48:13): [True: 0, False: 0] ------------------ 49| 0| return false; 50| 0| } 51| | 52| 0| pos += le + 1; 53| 0| } 54| 0| return true; 55| 46|} array_bitset_container_intersection: 21| 585| array_container_t *dst) { 22| 585| if (dst->capacity < src_1->cardinality) { ------------------ | Branch (22:9): [True: 585, False: 0] ------------------ 23| 585| array_container_grow(dst, src_1->cardinality, false); 24| 585| } 25| 585| int32_t newcard = 0; // dst could be src_1 26| 585| const int32_t origcard = src_1->cardinality; 27| 311k| for (int i = 0; i < origcard; ++i) { ------------------ | Branch (27:21): [True: 311k, False: 585] ------------------ 28| 311k| uint16_t key = src_1->array[i]; 29| | // this branchless approach is much faster... 30| 311k| dst->array[newcard] = key; 31| 311k| newcard += bitset_container_contains(src_2, key); 32| | /** 33| | * we could do it this way instead... 34| | * if (bitset_container_contains(src_2, key)) { 35| | * dst->array[newcard++] = key; 36| | * } 37| | * but if the result is unpredictible, the processor generates 38| | * many mispredicted branches. 39| | * Difference can be huge (from 3 cycles when predictible all the way 40| | * to 16 cycles when unpredictible. 41| | * See 42| | * https://github.com/lemire/Code-used-on-Daniel-Lemire-s-blog/blob/master/extra/bitset/c/arraybitsetintersection.c 43| | */ 44| 311k| } 45| 585| dst->cardinality = newcard; 46| 585|} array_bitset_container_intersection_cardinality: 50| 2.12k| const array_container_t *src_1, const bitset_container_t *src_2) { 51| 2.12k| int32_t newcard = 0; 52| 2.12k| const int32_t origcard = src_1->cardinality; 53| 365k| for (int i = 0; i < origcard; ++i) { ------------------ | Branch (53:21): [True: 363k, False: 2.12k] ------------------ 54| 363k| uint16_t key = src_1->array[i]; 55| 363k| newcard += bitset_container_contains(src_2, key); 56| 363k| } 57| 2.12k| return newcard; 58| 2.12k|} array_bitset_container_intersect: 61| 531| const bitset_container_t *src_2) { 62| 531| const int32_t origcard = src_1->cardinality; 63| 3.71k| for (int i = 0; i < origcard; ++i) { ------------------ | Branch (63:21): [True: 3.69k, False: 19] ------------------ 64| 3.69k| uint16_t key = src_1->array[i]; 65| 3.69k| if (bitset_container_contains(src_2, key)) return true; ------------------ | Branch (65:13): [True: 512, False: 3.18k] ------------------ 66| 3.69k| } 67| 19| return false; 68| 531|} array_run_container_intersection: 75| 3.18k| array_container_t *dst) { 76| 3.18k| if (run_container_is_full(src_2)) { ------------------ | Branch (76:9): [True: 24, False: 3.16k] ------------------ 77| 24| if (dst != src_1) array_container_copy(src_1, dst); ------------------ | Branch (77:13): [True: 24, False: 0] ------------------ 78| 24| return; 79| 24| } 80| 3.16k| if (dst->capacity < src_1->cardinality) { ------------------ | Branch (80:9): [True: 3.16k, False: 0] ------------------ 81| 3.16k| array_container_grow(dst, src_1->cardinality, false); 82| 3.16k| } 83| 3.16k| if (src_2->n_runs == 0) { ------------------ | Branch (83:9): [True: 0, False: 3.16k] ------------------ 84| 0| return; 85| 0| } 86| 3.16k| int32_t rlepos = 0; 87| 3.16k| int32_t arraypos = 0; 88| 3.16k| rle16_t rle = src_2->runs[rlepos]; 89| 3.16k| int32_t newcard = 0; 90| 419k| while (arraypos < src_1->cardinality) { ------------------ | Branch (90:12): [True: 417k, False: 2.60k] ------------------ 91| 417k| const uint16_t arrayval = src_1->array[arraypos]; 92| 564k| while (rle.value + rle.length < ------------------ | Branch (92:16): [True: 147k, False: 416k] ------------------ 93| 564k| arrayval) { // this will frequently be false 94| 147k| ++rlepos; 95| 147k| if (rlepos == src_2->n_runs) { ------------------ | Branch (95:17): [True: 556, False: 146k] ------------------ 96| 556| dst->cardinality = newcard; 97| 556| return; // we are done 98| 556| } 99| 146k| rle = src_2->runs[rlepos]; 100| 146k| } 101| 416k| if (rle.value > arrayval) { ------------------ | Branch (101:13): [True: 30.0k, False: 386k] ------------------ 102| 30.0k| arraypos = advanceUntil(src_1->array, arraypos, src_1->cardinality, 103| 30.0k| rle.value); 104| 386k| } else { 105| 386k| dst->array[newcard] = arrayval; 106| 386k| newcard++; 107| 386k| arraypos++; 108| 386k| } 109| 416k| } 110| 2.60k| dst->cardinality = newcard; 111| 2.60k|} run_bitset_container_intersection: 119| 613| container_t **dst) { 120| 613| if (run_container_is_full(src_1)) { ------------------ | Branch (120:9): [True: 0, False: 613] ------------------ 121| 0| if (*dst != src_2) *dst = bitset_container_clone(src_2); ------------------ | Branch (121:13): [True: 0, False: 0] ------------------ 122| 0| return true; 123| 0| } 124| 613| int32_t card = run_container_cardinality(src_1); 125| 613| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (125:9): [True: 613, False: 0] ------------------ 126| | // result can only be an array (assuming that we never make a 127| | // RunContainer) 128| 613| if (card > src_2->cardinality) { ------------------ | Branch (128:13): [True: 0, False: 613] ------------------ 129| 0| card = src_2->cardinality; 130| 0| } 131| 613| array_container_t *answer = array_container_create_given_capacity(card); 132| 613| *dst = answer; 133| 613| if (*dst == NULL) { ------------------ | Branch (133:13): [True: 0, False: 613] ------------------ 134| 0| return false; 135| 0| } 136| 51.0k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (136:34): [True: 50.4k, False: 613] ------------------ 137| 50.4k| rle16_t rle = src_1->runs[rlepos]; 138| 50.4k| uint32_t endofrun = (uint32_t)rle.value + rle.length; 139| 184k| for (uint32_t runValue = rle.value; runValue <= endofrun; ------------------ | Branch (139:49): [True: 134k, False: 50.4k] ------------------ 140| 134k| ++runValue) { 141| 134k| answer->array[answer->cardinality] = (uint16_t)runValue; 142| 134k| answer->cardinality += 143| 134k| bitset_container_contains(src_2, runValue); 144| 134k| } 145| 50.4k| } 146| 613| return false; 147| 613| } 148| 0| if (*dst == src_2) { // we attempt in-place ------------------ | Branch (148:9): [True: 0, False: 0] ------------------ 149| 0| bitset_container_t *answer = CAST_bitset(*dst); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 150| 0| uint32_t start = 0; 151| 0| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (151:34): [True: 0, False: 0] ------------------ 152| 0| const rle16_t rle = src_1->runs[rlepos]; 153| 0| uint32_t end = rle.value; 154| 0| bitset_reset_range(src_2->words, start, end); 155| | 156| 0| start = end + rle.length + 1; 157| 0| } 158| 0| bitset_reset_range(src_2->words, start, UINT32_C(1) << 16); 159| 0| answer->cardinality = bitset_container_compute_cardinality(answer); 160| 0| if (src_2->cardinality > DEFAULT_MAX_SIZE) { ------------------ | Branch (160:13): [True: 0, False: 0] ------------------ 161| 0| return true; 162| 0| } else { 163| 0| array_container_t *newanswer = array_container_from_bitset(src_2); 164| 0| if (newanswer == NULL) { ------------------ | Branch (164:17): [True: 0, False: 0] ------------------ 165| 0| *dst = NULL; 166| 0| return false; 167| 0| } 168| 0| *dst = newanswer; 169| 0| return false; 170| 0| } 171| 0| } else { // no inplace 172| | // we expect the answer to be a bitmap (if we are lucky) 173| 0| bitset_container_t *answer = bitset_container_clone(src_2); 174| | 175| 0| *dst = answer; 176| 0| if (answer == NULL) { ------------------ | Branch (176:13): [True: 0, False: 0] ------------------ 177| 0| return true; 178| 0| } 179| 0| uint32_t start = 0; 180| 0| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (180:34): [True: 0, False: 0] ------------------ 181| 0| const rle16_t rle = src_1->runs[rlepos]; 182| 0| uint32_t end = rle.value; 183| 0| bitset_reset_range(answer->words, start, end); 184| 0| start = end + rle.length + 1; 185| 0| } 186| 0| bitset_reset_range(answer->words, start, UINT32_C(1) << 16); 187| 0| answer->cardinality = bitset_container_compute_cardinality(answer); 188| | 189| 0| if (answer->cardinality > DEFAULT_MAX_SIZE) { ------------------ | Branch (189:13): [True: 0, False: 0] ------------------ 190| 0| return true; 191| 0| } else { 192| 0| array_container_t *newanswer = array_container_from_bitset(answer); 193| 0| bitset_container_free(CAST_bitset(*dst)); ------------------ | | 52| 0|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 194| 0| if (newanswer == NULL) { ------------------ | Branch (194:17): [True: 0, False: 0] ------------------ 195| 0| *dst = NULL; 196| 0| return false; 197| 0| } 198| 0| *dst = newanswer; 199| | return false; 200| 0| } 201| 0| } 202| 0|} array_run_container_intersection_cardinality: 206| 7.96k| const run_container_t *src_2) { 207| 7.96k| if (run_container_is_full(src_2)) { ------------------ | Branch (207:9): [True: 96, False: 7.86k] ------------------ 208| 96| return src_1->cardinality; 209| 96| } 210| 7.86k| if (src_2->n_runs == 0) { ------------------ | Branch (210:9): [True: 0, False: 7.86k] ------------------ 211| 0| return 0; 212| 0| } 213| 7.86k| int32_t rlepos = 0; 214| 7.86k| int32_t arraypos = 0; 215| 7.86k| rle16_t rle = src_2->runs[rlepos]; 216| 7.86k| int32_t newcard = 0; 217| 786k| while (arraypos < src_1->cardinality) { ------------------ | Branch (217:12): [True: 780k, False: 6.08k] ------------------ 218| 780k| const uint16_t arrayval = src_1->array[arraypos]; 219| 1.16M| while (rle.value + rle.length < ------------------ | Branch (219:16): [True: 383k, False: 778k] ------------------ 220| 1.16M| arrayval) { // this will frequently be false 221| 383k| ++rlepos; 222| 383k| if (rlepos == src_2->n_runs) { ------------------ | Branch (222:17): [True: 1.78k, False: 382k] ------------------ 223| 1.78k| return newcard; // we are done 224| 1.78k| } 225| 382k| rle = src_2->runs[rlepos]; 226| 382k| } 227| 778k| if (rle.value > arrayval) { ------------------ | Branch (227:13): [True: 106k, False: 671k] ------------------ 228| 106k| arraypos = advanceUntil(src_1->array, arraypos, src_1->cardinality, 229| 106k| rle.value); 230| 671k| } else { 231| 671k| newcard++; 232| 671k| arraypos++; 233| 671k| } 234| 778k| } 235| 6.08k| return newcard; 236| 7.86k|} run_bitset_container_intersection_cardinality: 241| 2.45k| const run_container_t *src_1, const bitset_container_t *src_2) { 242| 2.45k| if (run_container_is_full(src_1)) { ------------------ | Branch (242:9): [True: 0, False: 2.45k] ------------------ 243| 0| return bitset_container_cardinality(src_2); 244| 0| } 245| 2.45k| int answer = 0; 246| 204k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (246:30): [True: 201k, False: 2.45k] ------------------ 247| 201k| rle16_t rle = src_1->runs[rlepos]; 248| 201k| answer += 249| 201k| bitset_lenrange_cardinality(src_2->words, rle.value, rle.length); 250| 201k| } 251| 2.45k| return answer; 252| 2.45k|} array_run_container_intersect: 255| 1.99k| const run_container_t *src_2) { 256| 1.99k| if (run_container_is_full(src_2)) { ------------------ | Branch (256:9): [True: 24, False: 1.96k] ------------------ 257| 24| return !array_container_empty(src_1); 258| 24| } 259| 1.96k| if (src_2->n_runs == 0) { ------------------ | Branch (259:9): [True: 0, False: 1.96k] ------------------ 260| 0| return false; 261| 0| } 262| 1.96k| int32_t rlepos = 0; 263| 1.96k| int32_t arraypos = 0; 264| 1.96k| rle16_t rle = src_2->runs[rlepos]; 265| 3.51k| while (arraypos < src_1->cardinality) { ------------------ | Branch (265:12): [True: 3.45k, False: 61] ------------------ 266| 3.45k| const uint16_t arrayval = src_1->array[arraypos]; 267| 8.15k| while (rle.value + rle.length < ------------------ | Branch (267:16): [True: 4.74k, False: 3.41k] ------------------ 268| 8.15k| arrayval) { // this will frequently be false 269| 4.74k| ++rlepos; 270| 4.74k| if (rlepos == src_2->n_runs) { ------------------ | Branch (270:17): [True: 43, False: 4.70k] ------------------ 271| 43| return false; // we are done 272| 43| } 273| 4.70k| rle = src_2->runs[rlepos]; 274| 4.70k| } 275| 3.41k| if (rle.value > arrayval) { ------------------ | Branch (275:13): [True: 1.54k, False: 1.86k] ------------------ 276| 1.54k| arraypos = advanceUntil(src_1->array, arraypos, src_1->cardinality, 277| 1.54k| rle.value); 278| 1.86k| } else { 279| 1.86k| return true; 280| 1.86k| } 281| 3.41k| } 282| 61| return false; 283| 1.96k|} run_bitset_container_intersect: 288| 613| const bitset_container_t *src_2) { 289| 613| if (run_container_is_full(src_1)) { ------------------ | Branch (289:9): [True: 0, False: 613] ------------------ 290| 0| return !bitset_container_empty(src_2); 291| 0| } 292| 3.75k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (292:30): [True: 3.74k, False: 11] ------------------ 293| 3.74k| rle16_t rle = src_1->runs[rlepos]; 294| 3.74k| if (!bitset_lenrange_empty(src_2->words, rle.value, rle.length)) ------------------ | Branch (294:13): [True: 602, False: 3.13k] ------------------ 295| 602| return true; 296| 3.74k| } 297| 11| return false; 298| 613|} bitset_bitset_container_intersection_inplace: 329| 1.40k| container_t **dst) { 330| 1.40k| const int newCardinality = bitset_container_and_justcard(src_1, src_2); 331| 1.40k| if (newCardinality > DEFAULT_MAX_SIZE) { ------------------ | Branch (331:9): [True: 1.40k, False: 0] ------------------ 332| 1.40k| *dst = src_1; 333| 1.40k| bitset_container_and_nocard(src_1, src_2, src_1); 334| 1.40k| CAST_bitset(*dst)->cardinality = newCardinality; ------------------ | | 52| 1.40k|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 1.40k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 335| 1.40k| return true; // it is a bitset 336| 1.40k| } 337| 0| *dst = array_container_create_given_capacity(newCardinality); 338| 0| if (*dst != NULL) { ------------------ | Branch (338:9): [True: 0, False: 0] ------------------ 339| 0| CAST_array(*dst)->cardinality = newCardinality; ------------------ | | 54| 0|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 340| 0| bitset_extract_intersection_setbits_uint16( 341| 0| src_1->words, src_2->words, BITSET_CONTAINER_SIZE_IN_WORDS, 342| 0| CAST_array(*dst)->array, 0); ------------------ | | 54| 0|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 343| 0| } 344| | return false; // not a bitset 345| 1.40k|} array_container_negation: 33| 155| bitset_container_t *dst) { 34| 155| uint64_t card = UINT64_C(1 << 16); 35| 155| bitset_container_set_all(dst); 36| | 37| 155| if (src->cardinality == 0) { ------------------ | Branch (37:9): [True: 0, False: 155] ------------------ 38| 0| return; 39| 0| } 40| | 41| 155| dst->cardinality = (int32_t)bitset_clear_list(dst->words, card, src->array, 42| 155| (uint64_t)src->cardinality); 43| 155|} bitset_container_negation_inplace: 67| 61| container_t **dst) { 68| 61| return bitset_container_negation_range_inplace(src, 0, (1 << 16), dst); 69| 61|} run_container_negation_inplace: 88| 6.63k|int run_container_negation_inplace(run_container_t *src, container_t **dst) { 89| 6.63k| return run_container_negation_range_inplace(src, 0, (1 << 16), dst); 90| 6.63k|} array_container_negation_range: 99| 3.49k| container_t **dst) { 100| | /* close port of the Java implementation */ 101| 3.49k| if (range_start >= range_end) { ------------------ | Branch (101:9): [True: 0, False: 3.49k] ------------------ 102| 0| *dst = array_container_clone(src); 103| 0| return false; 104| 0| } 105| | 106| 3.49k| int32_t start_index = 107| 3.49k| binarySearch(src->array, src->cardinality, (uint16_t)range_start); 108| 3.49k| if (start_index < 0) start_index = -start_index - 1; ------------------ | Branch (108:9): [True: 3.23k, False: 267] ------------------ 109| | 110| 3.49k| int32_t last_index = 111| 3.49k| binarySearch(src->array, src->cardinality, (uint16_t)(range_end - 1)); 112| 3.49k| if (last_index < 0) last_index = -last_index - 2; ------------------ | Branch (112:9): [True: 3.26k, False: 233] ------------------ 113| | 114| 3.49k| const int32_t current_values_in_range = last_index - start_index + 1; 115| 3.49k| const int32_t span_to_be_flipped = range_end - range_start; 116| 3.49k| const int32_t new_values_in_range = 117| 3.49k| span_to_be_flipped - current_values_in_range; 118| 3.49k| const int32_t cardinality_change = 119| 3.49k| new_values_in_range - current_values_in_range; 120| 3.49k| const int32_t new_cardinality = src->cardinality + cardinality_change; 121| | 122| 3.49k| if (new_cardinality > DEFAULT_MAX_SIZE) { ------------------ | Branch (122:9): [True: 130, False: 3.36k] ------------------ 123| 130| bitset_container_t *temp = bitset_container_from_array(src); 124| 130| bitset_flip_range(temp->words, (uint32_t)range_start, 125| 130| (uint32_t)range_end); 126| 130| temp->cardinality = new_cardinality; 127| 130| *dst = temp; 128| 130| return true; 129| 130| } 130| | 131| 3.36k| array_container_t *arr = 132| 3.36k| array_container_create_given_capacity(new_cardinality); 133| 3.36k| *dst = (container_t *)arr; 134| 3.36k| if (new_cardinality == 0) { ------------------ | Branch (134:9): [True: 11, False: 3.35k] ------------------ 135| 11| arr->cardinality = new_cardinality; 136| 11| return false; // we are done. 137| 11| } 138| | // copy stuff before the active area 139| 3.35k| memcpy(arr->array, src->array, start_index * sizeof(uint16_t)); 140| | 141| | // work on the range 142| 3.35k| int32_t out_pos = start_index, in_pos = start_index; 143| 3.35k| int32_t val_in_range = range_start; 144| 47.5k| for (; val_in_range < range_end && in_pos <= last_index; ++val_in_range) { ------------------ | Branch (144:12): [True: 47.3k, False: 215] | Branch (144:40): [True: 44.1k, False: 3.14k] ------------------ 145| 44.1k| if ((uint16_t)val_in_range != src->array[in_pos]) { ------------------ | Branch (145:13): [True: 26.8k, False: 17.3k] ------------------ 146| 26.8k| arr->array[out_pos++] = (uint16_t)val_in_range; 147| 26.8k| } else { 148| 17.3k| ++in_pos; 149| 17.3k| } 150| 44.1k| } 151| 90.1k| for (; val_in_range < range_end; ++val_in_range) ------------------ | Branch (151:12): [True: 86.7k, False: 3.35k] ------------------ 152| 86.7k| arr->array[out_pos++] = (uint16_t)val_in_range; 153| | 154| | // content after the active range 155| 3.35k| memcpy(arr->array + out_pos, src->array + (last_index + 1), 156| 3.35k| (src->cardinality - (last_index + 1)) * sizeof(uint16_t)); 157| 3.35k| arr->cardinality = new_cardinality; 158| | return false; 159| 3.36k|} array_container_negation_range_inplace: 168| 3.49k| container_t **dst) { 169| 3.49k| bool ans = array_container_negation_range(src, range_start, range_end, dst); 170| | // TODO : try a real inplace version 171| 3.49k| array_container_free(src); 172| 3.49k| return ans; 173| 3.49k|} bitset_container_negation_range_inplace: 217| 1.29k| container_t **dst) { 218| 1.29k| bitset_flip_range(src->words, (uint32_t)range_start, (uint32_t)range_end); 219| 1.29k| src->cardinality = bitset_container_compute_cardinality(src); 220| 1.29k| if (src->cardinality > DEFAULT_MAX_SIZE) { ------------------ | Branch (220:9): [True: 1.22k, False: 71] ------------------ 221| 1.22k| *dst = src; 222| 1.22k| return true; 223| 1.22k| } 224| 71| *dst = array_container_from_bitset(src); 225| 71| bitset_container_free(src); 226| | return false; 227| 1.29k|} run_container_negation_range: 237| 181| container_t **dst) { 238| 181| uint8_t return_typecode; 239| | 240| | // follows the Java implementation 241| 181| if (range_end <= range_start) { ------------------ | Branch (241:9): [True: 0, False: 181] ------------------ 242| 0| *dst = run_container_clone(src); 243| 0| return RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 244| 0| } 245| | 246| 181| run_container_t *ans = run_container_create_given_capacity( 247| 181| src->n_runs + 1); // src->n_runs + 1); 248| 181| int k = 0; 249| 2.89k| for (; k < src->n_runs && src->runs[k].value < range_start; ++k) { ------------------ | Branch (249:12): [True: 2.85k, False: 48] | Branch (249:31): [True: 2.71k, False: 133] ------------------ 250| 2.71k| ans->runs[k] = src->runs[k]; 251| 2.71k| ans->n_runs++; 252| 2.71k| } 253| | 254| 181| run_container_smart_append_exclusive( 255| 181| ans, (uint16_t)range_start, (uint16_t)(range_end - range_start - 1)); 256| | 257| 4.84k| for (; k < src->n_runs; ++k) { ------------------ | Branch (257:12): [True: 4.66k, False: 181] ------------------ 258| 4.66k| run_container_smart_append_exclusive(ans, src->runs[k].value, 259| 4.66k| src->runs[k].length); 260| 4.66k| } 261| | 262| 181| *dst = convert_run_to_efficient_container(ans, &return_typecode); 263| 181| if (return_typecode != RUN_CONTAINER_TYPE) run_container_free(ans); ------------------ | | 50| 181|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (263:9): [True: 9, False: 172] ------------------ 264| | 265| 181| return return_typecode; 266| 181|} run_container_negation_range_inplace: 278| 8.34k| container_t **dst) { 279| 8.34k| uint8_t return_typecode; 280| | 281| 8.34k| if (range_end <= range_start) { ------------------ | Branch (281:9): [True: 0, False: 8.34k] ------------------ 282| 0| *dst = src; 283| 0| return RUN_CONTAINER_TYPE; ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 284| 0| } 285| | 286| | // TODO: efficient special case when range is 0 to 65535 inclusive 287| | 288| 8.34k| if (src->capacity == src->n_runs) { ------------------ | Branch (288:9): [True: 7.19k, False: 1.15k] ------------------ 289| | // no excess room. More checking to see if result can fit 290| 7.19k| bool last_val_before_range = false; 291| 7.19k| bool first_val_in_range = false; 292| 7.19k| bool last_val_in_range = false; 293| 7.19k| bool first_val_past_range = false; 294| | 295| 7.19k| if (range_start > 0) ------------------ | Branch (295:13): [True: 176, False: 7.01k] ------------------ 296| 176| last_val_before_range = 297| 176| run_container_contains(src, (uint16_t)(range_start - 1)); 298| 7.19k| first_val_in_range = run_container_contains(src, (uint16_t)range_start); 299| | 300| 7.19k| if (last_val_before_range == first_val_in_range) { ------------------ | Branch (300:13): [True: 505, False: 6.68k] ------------------ 301| 505| last_val_in_range = 302| 505| run_container_contains(src, (uint16_t)(range_end - 1)); 303| 505| if (range_end != 0x10000) ------------------ | Branch (303:17): [True: 313, False: 192] ------------------ 304| 313| first_val_past_range = 305| 313| run_container_contains(src, (uint16_t)range_end); 306| | 307| 505| if (last_val_in_range == ------------------ | Branch (307:17): [True: 181, False: 324] ------------------ 308| 505| first_val_past_range) { // no space for inplace 309| 181| int ans = run_container_negation_range(src, range_start, 310| 181| range_end, dst); 311| 181| run_container_free(src); 312| 181| return ans; 313| 181| } 314| 505| } 315| 7.19k| } 316| | // all other cases: result will fit 317| | 318| 8.16k| run_container_t *ans = src; 319| 8.16k| int my_nbr_runs = src->n_runs; 320| | 321| 8.16k| ans->n_runs = 0; 322| 8.16k| int k = 0; 323| 10.5k| for (; (k < my_nbr_runs) && (src->runs[k].value < range_start); ++k) { ------------------ | Branch (323:12): [True: 10.4k, False: 75] | Branch (323:33): [True: 2.36k, False: 8.08k] ------------------ 324| | // ans->runs[k] = src->runs[k]; (would be self-copy) 325| 2.36k| ans->n_runs++; 326| 2.36k| } 327| | 328| | // as with Java implementation, use locals to give self a buffer of depth 1 329| 8.16k| rle16_t buffered = CROARING_MAKE_RLE16(0, 0); ------------------ | | 60| 8.16k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 330| 8.16k| rle16_t next = buffered; 331| 8.16k| if (k < my_nbr_runs) buffered = src->runs[k]; ------------------ | Branch (331:9): [True: 8.08k, False: 75] ------------------ 332| | 333| 8.16k| run_container_smart_append_exclusive( 334| 8.16k| ans, (uint16_t)range_start, (uint16_t)(range_end - range_start - 1)); 335| | 336| 95.5k| for (; k < my_nbr_runs; ++k) { ------------------ | Branch (336:12): [True: 87.4k, False: 8.16k] ------------------ 337| 87.4k| if (k + 1 < my_nbr_runs) next = src->runs[k + 1]; ------------------ | Branch (337:13): [True: 79.3k, False: 8.08k] ------------------ 338| | 339| 87.4k| run_container_smart_append_exclusive(ans, buffered.value, 340| 87.4k| buffered.length); 341| 87.4k| buffered = next; 342| 87.4k| } 343| | 344| 8.16k| *dst = convert_run_to_efficient_container(ans, &return_typecode); 345| 8.16k| if (return_typecode != RUN_CONTAINER_TYPE) run_container_free(ans); ------------------ | | 50| 8.16k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (345:9): [True: 6.47k, False: 1.68k] ------------------ 346| | 347| 8.16k| return return_typecode; 348| 8.34k|} array_container_is_subset_bitset: 11| 2.20k| const bitset_container_t* container2) { 12| 2.20k| if (container2->cardinality != BITSET_UNKNOWN_CARDINALITY) { ------------------ | Branch (12:9): [True: 2.20k, False: 0] ------------------ 13| 2.20k| if (container2->cardinality < container1->cardinality) { ------------------ | Branch (13:13): [True: 0, False: 2.20k] ------------------ 14| 0| return false; 15| 0| } 16| 2.20k| } 17| 2.20k| for (int i = 0; i < container1->cardinality; ++i) { ------------------ | Branch (17:21): [True: 2.20k, False: 0] ------------------ 18| 2.20k| if (!bitset_container_contains(container2, container1->array[i])) { ------------------ | Branch (18:13): [True: 2.20k, False: 0] ------------------ 19| 2.20k| return false; 20| 2.20k| } 21| 2.20k| } 22| 0| return true; 23| 2.20k|} run_container_is_subset_array: 26| 391| const array_container_t* container2) { 27| 391| if (run_container_cardinality(container1) > container2->cardinality) ------------------ | Branch (27:9): [True: 173, False: 218] ------------------ 28| 173| return false; 29| 218| int32_t start_pos = -1, stop_pos = -1; 30| 218| for (int i = 0; i < container1->n_runs; ++i) { ------------------ | Branch (30:21): [True: 218, False: 0] ------------------ 31| 218| int32_t start = container1->runs[i].value; 32| 218| int32_t stop = start + container1->runs[i].length; 33| 218| start_pos = advanceUntil(container2->array, stop_pos, 34| 218| container2->cardinality, start); 35| 218| stop_pos = advanceUntil(container2->array, stop_pos, 36| 218| container2->cardinality, stop); 37| 218| if (stop_pos == container2->cardinality) { ------------------ | Branch (37:13): [True: 5, False: 213] ------------------ 38| 5| return false; 39| 213| } else if (stop_pos - start_pos != stop - start || ------------------ | Branch (39:20): [True: 78, False: 135] ------------------ 40| 135| container2->array[start_pos] != start || ------------------ | Branch (40:20): [True: 135, False: 0] ------------------ 41| 213| container2->array[stop_pos] != stop) { ------------------ | Branch (41:20): [True: 0, False: 0] ------------------ 42| 213| return false; 43| 213| } 44| 218| } 45| 0| return true; 46| 218|} array_container_is_subset_run: 49| 2.27k| const run_container_t* container2) { 50| 2.27k| if (container1->cardinality > run_container_cardinality(container2)) ------------------ | Branch (50:9): [True: 179, False: 2.09k] ------------------ 51| 179| return false; 52| 2.09k| int i_array = 0, i_run = 0; 53| 5.48k| while (i_array < container1->cardinality && i_run < container2->n_runs) { ------------------ | Branch (53:12): [True: 5.48k, False: 0] | Branch (53:49): [True: 5.47k, False: 13] ------------------ 54| 5.47k| uint32_t start = container2->runs[i_run].value; 55| 5.47k| uint32_t stop = start + container2->runs[i_run].length; 56| 5.47k| if (container1->array[i_array] < start) { ------------------ | Branch (56:13): [True: 2.08k, False: 3.38k] ------------------ 57| 2.08k| return false; 58| 3.38k| } else if (container1->array[i_array] > stop) { ------------------ | Branch (58:20): [True: 3.38k, False: 0] ------------------ 59| 3.38k| i_run++; 60| 3.38k| } else { // the value of the array is in the run 61| 0| i_array++; 62| 0| } 63| 5.47k| } 64| 13| if (i_array == container1->cardinality) { ------------------ | Branch (64:9): [True: 0, False: 13] ------------------ 65| 0| return true; 66| 13| } else { 67| | return false; 68| 13| } 69| 13|} array_bitset_container_union: 24| 1.11k| bitset_container_t *dst) { 25| 1.11k| if (src_2 != dst) bitset_container_copy(src_2, dst); ------------------ | Branch (25:9): [True: 1.11k, False: 0] ------------------ 26| 1.11k| dst->cardinality = (int32_t)bitset_set_list_withcard( 27| 1.11k| dst->words, dst->cardinality, src_1->array, src_1->cardinality); 28| 1.11k|} run_bitset_container_union: 43| 1.22k| bitset_container_t *dst) { 44| 1.22k| assert(!run_container_is_full(src_1)); // catch this case upstream 45| 1.22k| if (src_2 != dst) bitset_container_copy(src_2, dst); ------------------ | Branch (45:9): [True: 1.22k, False: 0] ------------------ 46| 102k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (46:30): [True: 100k, False: 1.22k] ------------------ 47| 100k| rle16_t rle = src_1->runs[rlepos]; 48| 100k| bitset_set_lenrange(dst->words, rle.value, rle.length); 49| 100k| } 50| 1.22k| dst->cardinality = bitset_container_compute_cardinality(dst); 51| 1.22k|} array_run_container_union: 68| 2.62k| run_container_t *dst) { 69| 2.62k| if (run_container_is_full(src_2)) { ------------------ | Branch (69:9): [True: 93, False: 2.53k] ------------------ 70| 93| run_container_copy(src_2, dst); 71| 93| return; 72| 93| } 73| | // TODO: see whether the "2*" is spurious 74| 2.53k| run_container_grow(dst, 2 * (src_1->cardinality + src_2->n_runs), false); 75| 2.53k| int32_t rlepos = 0; 76| 2.53k| int32_t arraypos = 0; 77| 2.53k| rle16_t previousrle; 78| 2.53k| if (src_2->runs[rlepos].value <= src_1->array[arraypos]) { ------------------ | Branch (78:9): [True: 1.86k, False: 670] ------------------ 79| 1.86k| previousrle = run_container_append_first(dst, src_2->runs[rlepos]); 80| 1.86k| rlepos++; 81| 1.86k| } else { 82| 670| previousrle = 83| 670| run_container_append_value_first(dst, src_1->array[arraypos]); 84| 670| arraypos++; 85| 670| } 86| 466k| while ((rlepos < src_2->n_runs) && (arraypos < src_1->cardinality)) { ------------------ | Branch (86:12): [True: 464k, False: 1.61k] | Branch (86:40): [True: 463k, False: 915] ------------------ 87| 463k| if (src_2->runs[rlepos].value <= src_1->array[arraypos]) { ------------------ | Branch (87:13): [True: 150k, False: 313k] ------------------ 88| 150k| run_container_append(dst, src_2->runs[rlepos], &previousrle); 89| 150k| rlepos++; 90| 313k| } else { 91| 313k| run_container_append_value(dst, src_1->array[arraypos], 92| 313k| &previousrle); 93| 313k| arraypos++; 94| 313k| } 95| 463k| } 96| 2.53k| if (arraypos < src_1->cardinality) { ------------------ | Branch (96:9): [True: 1.61k, False: 915] ------------------ 97| 147k| while (arraypos < src_1->cardinality) { ------------------ | Branch (97:16): [True: 146k, False: 1.61k] ------------------ 98| 146k| run_container_append_value(dst, src_1->array[arraypos], 99| 146k| &previousrle); 100| 146k| arraypos++; 101| 146k| } 102| 1.61k| } else { 103| 6.81k| while (rlepos < src_2->n_runs) { ------------------ | Branch (103:16): [True: 5.89k, False: 915] ------------------ 104| 5.89k| run_container_append(dst, src_2->runs[rlepos], &previousrle); 105| 5.89k| rlepos++; 106| 5.89k| } 107| 915| } 108| 2.53k|} array_run_container_inplace_union: 111| 259| run_container_t *src_2) { 112| 259| if (run_container_is_full(src_2)) { ------------------ | Branch (112:9): [True: 0, False: 259] ------------------ 113| 0| return; 114| 0| } 115| 259| const int32_t maxoutput = src_1->cardinality + src_2->n_runs; 116| 259| const int32_t neededcapacity = maxoutput + src_2->n_runs; 117| 259| if (src_2->capacity < neededcapacity) ------------------ | Branch (117:9): [True: 259, False: 0] ------------------ 118| 259| run_container_grow(src_2, neededcapacity, true); 119| 259| memmove(src_2->runs + maxoutput, src_2->runs, 120| 259| src_2->n_runs * sizeof(rle16_t)); 121| 259| rle16_t *inputsrc2 = src_2->runs + maxoutput; 122| 259| int32_t rlepos = 0; 123| 259| int32_t arraypos = 0; 124| 259| int src2nruns = src_2->n_runs; 125| 259| src_2->n_runs = 0; 126| | 127| 259| rle16_t previousrle; 128| | 129| 259| if (inputsrc2[rlepos].value <= src_1->array[arraypos]) { ------------------ | Branch (129:9): [True: 224, False: 35] ------------------ 130| 224| previousrle = run_container_append_first(src_2, inputsrc2[rlepos]); 131| 224| rlepos++; 132| 224| } else { 133| 35| previousrle = 134| 35| run_container_append_value_first(src_2, src_1->array[arraypos]); 135| 35| arraypos++; 136| 35| } 137| | 138| 54.3k| while ((rlepos < src2nruns) && (arraypos < src_1->cardinality)) { ------------------ | Branch (138:12): [True: 54.1k, False: 259] | Branch (138:36): [True: 54.1k, False: 0] ------------------ 139| 54.1k| if (inputsrc2[rlepos].value <= src_1->array[arraypos]) { ------------------ | Branch (139:13): [True: 14.2k, False: 39.8k] ------------------ 140| 14.2k| run_container_append(src_2, inputsrc2[rlepos], &previousrle); 141| 14.2k| rlepos++; 142| 39.8k| } else { 143| 39.8k| run_container_append_value(src_2, src_1->array[arraypos], 144| 39.8k| &previousrle); 145| 39.8k| arraypos++; 146| 39.8k| } 147| 54.1k| } 148| 259| if (arraypos < src_1->cardinality) { ------------------ | Branch (148:9): [True: 259, False: 0] ------------------ 149| 2.70k| while (arraypos < src_1->cardinality) { ------------------ | Branch (149:16): [True: 2.44k, False: 259] ------------------ 150| 2.44k| run_container_append_value(src_2, src_1->array[arraypos], 151| 2.44k| &previousrle); 152| 2.44k| arraypos++; 153| 2.44k| } 154| 259| } else { 155| 0| while (rlepos < src2nruns) { ------------------ | Branch (155:16): [True: 0, False: 0] ------------------ 156| 0| run_container_append(src_2, inputsrc2[rlepos], &previousrle); 157| 0| rlepos++; 158| 0| } 159| 0| } 160| 259|} array_array_container_union: 164| 2.30k| container_t **dst) { 165| 2.30k| int totalCardinality = src_1->cardinality + src_2->cardinality; 166| 2.30k| if (totalCardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (166:9): [True: 2.18k, False: 116] ------------------ 167| 2.18k| *dst = array_container_create_given_capacity(totalCardinality); 168| 2.18k| if (*dst != NULL) { ------------------ | Branch (168:13): [True: 2.18k, False: 0] ------------------ 169| 2.18k| array_container_union(src_1, src_2, CAST_array(*dst)); ------------------ | | 54| 2.18k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.18k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 170| 2.18k| } else { 171| 0| return true; // otherwise failure won't be caught 172| 0| } 173| 2.18k| return false; // not a bitset 174| 2.18k| } 175| 116| *dst = bitset_container_create(); 176| 116| bool returnval = true; // expect a bitset 177| 116| if (*dst != NULL) { ------------------ | Branch (177:9): [True: 116, False: 0] ------------------ 178| 116| bitset_container_t *ourbitset = CAST_bitset(*dst); ------------------ | | 52| 116|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 116|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 179| 116| bitset_set_list(ourbitset->words, src_1->array, src_1->cardinality); 180| 116| ourbitset->cardinality = (int32_t)bitset_set_list_withcard( 181| 116| ourbitset->words, src_1->cardinality, src_2->array, 182| 116| src_2->cardinality); 183| 116| if (ourbitset->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (183:13): [True: 62, False: 54] ------------------ 184| | // need to convert! 185| 62| *dst = array_container_from_bitset(ourbitset); 186| 62| bitset_container_free(ourbitset); 187| | returnval = false; // not going to be a bitset 188| 62| } 189| 116| } 190| 116| return returnval; 191| 2.30k|} array_array_container_inplace_union: 195| 2.47k| container_t **dst) { 196| 2.47k| int totalCardinality = src_1->cardinality + src_2->cardinality; 197| 2.47k| *dst = NULL; 198| 2.47k| if (totalCardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (198:9): [True: 2.40k, False: 77] ------------------ 199| 2.40k| if (src_1->capacity < totalCardinality) { ------------------ | Branch (199:13): [True: 2.40k, False: 0] ------------------ 200| 2.40k| *dst = array_container_create_given_capacity( 201| 2.40k| 2 * totalCardinality); // be purposefully generous 202| 2.40k| if (*dst != NULL) { ------------------ | Branch (202:17): [True: 2.40k, False: 0] ------------------ 203| 2.40k| array_container_union(src_1, src_2, CAST_array(*dst)); ------------------ | | 54| 2.40k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 2.40k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 204| 2.40k| } else { 205| 0| return true; // otherwise failure won't be caught 206| 0| } 207| 2.40k| return false; // not a bitset 208| 2.40k| } else { 209| 0| memmove(src_1->array + src_2->cardinality, src_1->array, 210| 0| src_1->cardinality * sizeof(uint16_t)); 211| | // In theory, we could use fast_union_uint16, but it is unsafe. It 212| | // fails with Intel compilers in particular. 213| | // https://github.com/RoaringBitmap/CRoaring/pull/452 214| | // See report https://github.com/RoaringBitmap/CRoaring/issues/476 215| 0| src_1->cardinality = (int32_t)union_uint16( 216| 0| src_1->array + src_2->cardinality, src_1->cardinality, 217| 0| src_2->array, src_2->cardinality, src_1->array); 218| 0| return false; // not a bitset 219| 0| } 220| 2.40k| } 221| 77| *dst = bitset_container_create(); 222| 77| bool returnval = true; // expect a bitset 223| 77| if (*dst != NULL) { ------------------ | Branch (223:9): [True: 77, False: 0] ------------------ 224| 77| bitset_container_t *ourbitset = CAST_bitset(*dst); ------------------ | | 52| 77|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 77|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 225| 77| bitset_set_list(ourbitset->words, src_1->array, src_1->cardinality); 226| 77| ourbitset->cardinality = (int32_t)bitset_set_list_withcard( 227| 77| ourbitset->words, src_1->cardinality, src_2->array, 228| 77| src_2->cardinality); 229| 77| if (ourbitset->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (229:13): [True: 77, False: 0] ------------------ 230| | // need to convert! 231| 77| if (src_1->capacity < ourbitset->cardinality) { ------------------ | Branch (231:17): [True: 77, False: 0] ------------------ 232| 77| array_container_grow(src_1, ourbitset->cardinality, false); 233| 77| } 234| | 235| 77| bitset_extract_setbits_uint16(ourbitset->words, 236| 77| BITSET_CONTAINER_SIZE_IN_WORDS, 237| 77| src_1->array, 0); 238| 77| src_1->cardinality = ourbitset->cardinality; 239| 77| *dst = src_1; 240| 77| bitset_container_free(ourbitset); 241| | returnval = false; // not going to be a bitset 242| 77| } 243| 77| } 244| 77| return returnval; 245| 2.47k|} array_bitset_container_xor: 25| 619| container_t **dst) { 26| 619| bitset_container_t *result = bitset_container_create(); 27| 619| bitset_container_copy(src_2, result); 28| 619| result->cardinality = (int32_t)bitset_flip_list_withcard( 29| 619| result->words, result->cardinality, src_1->array, src_1->cardinality); 30| | 31| | // do required type conversions. 32| 619| if (result->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (32:9): [True: 136, False: 483] ------------------ 33| 136| *dst = array_container_from_bitset(result); 34| 136| bitset_container_free(result); 35| 136| return false; // not bitset 36| 136| } 37| 483| *dst = result; 38| | return true; // bitset 39| 619|} run_bitset_container_xor: 63| 698| container_t **dst) { 64| 698| bitset_container_t *result = bitset_container_create(); 65| | 66| 698| bitset_container_copy(src_2, result); 67| 54.5k| for (int32_t rlepos = 0; rlepos < src_1->n_runs; ++rlepos) { ------------------ | Branch (67:30): [True: 53.8k, False: 698] ------------------ 68| 53.8k| rle16_t rle = src_1->runs[rlepos]; 69| 53.8k| bitset_flip_range(result->words, rle.value, 70| 53.8k| rle.value + rle.length + UINT32_C(1)); 71| 53.8k| } 72| 698| result->cardinality = bitset_container_compute_cardinality(result); 73| | 74| 698| if (result->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (74:9): [True: 134, False: 564] ------------------ 75| 134| *dst = array_container_from_bitset(result); 76| 134| bitset_container_free(result); 77| 134| return false; // not bitset 78| 134| } 79| 564| *dst = result; 80| | return true; // bitset 81| 698|} array_run_container_xor: 105| 2.31k| const run_container_t *src_2, container_t **dst) { 106| | // semi following Java XOR implementation as of May 2016 107| | // the C OR implementation works quite differently and can return a run 108| | // container 109| | // TODO could optimize for full run containers. 110| | 111| | // use of lazy following Java impl. 112| 2.31k| const int arbitrary_threshold = 32; 113| 2.31k| if (src_1->cardinality < arbitrary_threshold) { ------------------ | Branch (113:9): [True: 893, False: 1.41k] ------------------ 114| 893| run_container_t *ans = run_container_create(); 115| 893| array_run_container_lazy_xor(src_1, src_2, ans); // keeps runs. 116| 893| uint8_t typecode_after; 117| 893| *dst = 118| 893| convert_run_to_efficient_container_and_free(ans, &typecode_after); 119| 893| return typecode_after; 120| 893| } 121| | 122| 1.41k| int card = run_container_cardinality(src_2); 123| 1.41k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (123:9): [True: 1.15k, False: 266] ------------------ 124| | // Java implementation works with the array, xoring the run elements via 125| | // iterator 126| 1.15k| array_container_t *temp = array_container_from_run(src_2); 127| 1.15k| bool ret_is_bitset = array_array_container_xor(temp, src_1, dst); 128| 1.15k| array_container_free(temp); 129| 1.15k| return ret_is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE; ------------------ | | 48| 3|#define BITSET_CONTAINER_TYPE 1 ------------------ return ret_is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE; ------------------ | | 49| 1.14k|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (129:16): [True: 3, False: 1.14k] ------------------ 130| | 131| 1.15k| } else { // guess that it will end up as a bitset 132| 266| bitset_container_t *result = bitset_container_from_run(src_2); 133| 266| bool is_bitset = bitset_array_container_ixor(result, src_1, dst); 134| | // any necessary type conversion has been done by the ixor 135| 266| int retval = (is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE); ------------------ | | 48| 247|#define BITSET_CONTAINER_TYPE 1 ------------------ int retval = (is_bitset ? BITSET_CONTAINER_TYPE : ARRAY_CONTAINER_TYPE); ------------------ | | 49| 19|#define ARRAY_CONTAINER_TYPE 2 ------------------ | Branch (135:23): [True: 247, False: 19] ------------------ 136| 266| return retval; 137| 266| } 138| 1.41k|} array_run_container_lazy_xor: 147| 893| run_container_t *dst) { 148| 893| run_container_grow(dst, src_1->cardinality + src_2->n_runs, false); 149| 893| int32_t rlepos = 0; 150| 893| int32_t arraypos = 0; 151| 893| dst->n_runs = 0; 152| | 153| 27.8k| while ((rlepos < src_2->n_runs) && (arraypos < src_1->cardinality)) { ------------------ | Branch (153:12): [True: 27.2k, False: 547] | Branch (153:40): [True: 26.9k, False: 346] ------------------ 154| 26.9k| if (src_2->runs[rlepos].value <= src_1->array[arraypos]) { ------------------ | Branch (154:13): [True: 16.3k, False: 10.6k] ------------------ 155| 16.3k| run_container_smart_append_exclusive(dst, src_2->runs[rlepos].value, 156| 16.3k| src_2->runs[rlepos].length); 157| 16.3k| rlepos++; 158| 16.3k| } else { 159| 10.6k| run_container_smart_append_exclusive(dst, src_1->array[arraypos], 160| 10.6k| 0); 161| 10.6k| arraypos++; 162| 10.6k| } 163| 26.9k| } 164| 3.06k| while (arraypos < src_1->cardinality) { ------------------ | Branch (164:12): [True: 2.17k, False: 893] ------------------ 165| 2.17k| run_container_smart_append_exclusive(dst, src_1->array[arraypos], 0); 166| 2.17k| arraypos++; 167| 2.17k| } 168| 9.05k| while (rlepos < src_2->n_runs) { ------------------ | Branch (168:12): [True: 8.16k, False: 893] ------------------ 169| 8.16k| run_container_smart_append_exclusive(dst, src_2->runs[rlepos].value, 170| 8.16k| src_2->runs[rlepos].length); 171| 8.16k| rlepos++; 172| 8.16k| } 173| 893|} run_run_container_xor: 180| 55.7k| const run_container_t *src_2, container_t **dst) { 181| 55.7k| run_container_t *ans = run_container_create(); 182| 55.7k| run_container_xor(src_1, src_2, ans); 183| 55.7k| uint8_t typecode_after; 184| 55.7k| *dst = convert_run_to_efficient_container_and_free(ans, &typecode_after); 185| 55.7k| return typecode_after; 186| 55.7k|} array_array_container_xor: 198| 8.23k| container_t **dst) { 199| 8.23k| int totalCardinality = 200| 8.23k| src_1->cardinality + src_2->cardinality; // upper bound 201| 8.23k| if (totalCardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (201:9): [True: 7.71k, False: 519] ------------------ 202| 7.71k| *dst = array_container_create_given_capacity(totalCardinality); 203| 7.71k| array_container_xor(src_1, src_2, CAST_array(*dst)); ------------------ | | 54| 7.71k|#define CAST_array(c) CAST(array_container_t *, c) // safer downcast | | ------------------ | | | | 86| 7.71k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 204| 7.71k| return false; // not a bitset 205| 7.71k| } 206| 519| *dst = bitset_container_from_array(src_1); 207| 519| bool returnval = true; // expect a bitset 208| 519| bitset_container_t *ourbitset = CAST_bitset(*dst); ------------------ | | 52| 519|#define CAST_bitset(c) CAST(bitset_container_t *, c) // safer downcast | | ------------------ | | | | 86| 519|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 209| 519| ourbitset->cardinality = (uint32_t)bitset_flip_list_withcard( 210| 519| ourbitset->words, src_1->cardinality, src_2->array, src_2->cardinality); 211| 519| if (ourbitset->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (211:9): [True: 504, False: 15] ------------------ 212| | // need to convert! 213| 504| *dst = array_container_from_bitset(ourbitset); 214| 504| bitset_container_free(ourbitset); 215| 504| returnval = false; // not going to be a bitset 216| 504| } 217| | 218| 519| return returnval; 219| 8.23k|} bitset_array_container_ixor: 284| 266| container_t **dst) { 285| 266| *dst = src_1; 286| 266| src_1->cardinality = (uint32_t)bitset_flip_list_withcard( 287| 266| src_1->words, src_1->cardinality, src_2->array, src_2->cardinality); 288| | 289| 266| if (src_1->cardinality <= DEFAULT_MAX_SIZE) { ------------------ | Branch (289:9): [True: 19, False: 247] ------------------ 290| 19| *dst = array_container_from_bitset(src_1); 291| 19| bitset_container_free(src_1); 292| 19| return false; // not bitset 293| 19| } else 294| 247| return true; 295| 266|} bitset_bitset_container_ixor: 304| 1.45k| container_t **dst) { 305| 1.45k| int card = bitset_container_xor(src_1, src_2, src_1); 306| 1.45k| if (card <= DEFAULT_MAX_SIZE) { ------------------ | Branch (306:9): [True: 1.45k, False: 0] ------------------ 307| 1.45k| *dst = array_container_from_bitset(src_1); 308| 1.45k| bitset_container_free(src_1); 309| 1.45k| return false; // not bitset 310| 1.45k| } else { 311| 0| *dst = src_1; 312| | return true; 313| 0| } 314| 1.45k|} array_bitset_container_ixor: 318| 88| container_t **dst) { 319| | bool ans = array_bitset_container_xor(src_1, src_2, dst); 320| 88| array_container_free(src_1); 321| 88| return ans; 322| 88|} run_bitset_container_ixor: 333| 22| container_t **dst) { 334| | bool ans = run_bitset_container_xor(src_1, src_2, dst); 335| 22| run_container_free(src_1); 336| 22| return ans; 337| 22|} bitset_run_container_ixor: 341| 63| container_t **dst) { 342| | bool ans = run_bitset_container_xor(src_2, src_1, dst); 343| 63| bitset_container_free(src_1); 344| 63| return ans; 345| 63|} array_run_container_ixor: 352| 262| const run_container_t *src_2, container_t **dst) { 353| 262| int ans = array_run_container_xor(src_1, src_2, dst); 354| 262| array_container_free(src_1); 355| 262| return ans; 356| 262|} run_array_container_ixor: 360| 58| container_t **dst) { 361| 58| int ans = array_run_container_xor(src_2, src_1, dst); 362| 58| run_container_free(src_1); 363| 58| return ans; 364| 58|} array_array_container_ixor: 368| 4.78k| container_t **dst) { 369| | bool ans = array_array_container_xor(src_1, src_2, dst); 370| 4.78k| array_container_free(src_1); 371| 4.78k| return ans; 372| 4.78k|} run_run_container_ixor: 375| 55.0k| container_t **dst) { 376| 55.0k| int ans = run_run_container_xor(src_1, src_2, dst); 377| 55.0k| run_container_free(src_1); 378| 55.0k| return ans; 379| 55.0k|} run_container_add: 39| 412k|bool run_container_add(run_container_t *run, uint16_t pos) { 40| 412k| int32_t index = interleavedBinarySearch(run->runs, run->n_runs, pos); 41| 412k| if (index >= 0) return false; // already there ------------------ | Branch (41:9): [True: 179k, False: 232k] ------------------ 42| 232k| index = -index - 2; // points to preceding value, possibly -1 43| 232k| if (index >= 0) { // possible match ------------------ | Branch (43:9): [True: 232k, False: 458] ------------------ 44| 232k| int32_t offset = pos - run->runs[index].value; 45| 232k| int32_t le = run->runs[index].length; 46| 232k| if (offset <= le) return false; // already there ------------------ | Branch (46:13): [True: 138k, False: 93.6k] ------------------ 47| 93.6k| if (offset == le + 1) { ------------------ | Branch (47:13): [True: 14.4k, False: 79.2k] ------------------ 48| | // we may need to fuse 49| 14.4k| if (index + 1 < run->n_runs) { ------------------ | Branch (49:17): [True: 14.1k, False: 248] ------------------ 50| 14.1k| if (run->runs[index + 1].value == pos + 1) { ------------------ | Branch (50:21): [True: 6.59k, False: 7.59k] ------------------ 51| | // indeed fusion is needed 52| 6.59k| run->runs[index].length = run->runs[index + 1].value + 53| 6.59k| run->runs[index + 1].length - 54| 6.59k| run->runs[index].value; 55| 6.59k| recoverRoomAtIndex(run, (uint16_t)(index + 1)); 56| 6.59k| return true; 57| 6.59k| } 58| 14.1k| } 59| 7.84k| run->runs[index].length++; 60| 7.84k| return true; 61| 14.4k| } 62| 79.2k| if (index + 1 < run->n_runs) { ------------------ | Branch (62:13): [True: 77.3k, False: 1.86k] ------------------ 63| | // we may need to fuse 64| 77.3k| if (run->runs[index + 1].value == pos + 1) { ------------------ | Branch (64:17): [True: 33.6k, False: 43.7k] ------------------ 65| | // indeed fusion is needed 66| 33.6k| run->runs[index + 1].value = pos; 67| 33.6k| run->runs[index + 1].length = run->runs[index + 1].length + 1; 68| 33.6k| return true; 69| 33.6k| } 70| 77.3k| } 71| 79.2k| } 72| 46.0k| if (index == -1) { ------------------ | Branch (72:9): [True: 458, False: 45.5k] ------------------ 73| | // we may need to extend the first run 74| 458| if (0 < run->n_runs) { ------------------ | Branch (74:13): [True: 458, False: 0] ------------------ 75| 458| if (run->runs[0].value == pos + 1) { ------------------ | Branch (75:17): [True: 284, False: 174] ------------------ 76| 284| run->runs[0].length++; 77| 284| run->runs[0].value--; 78| 284| return true; 79| 284| } 80| 458| } 81| 458| } 82| 45.7k| makeRoomAtIndex(run, (uint16_t)(index + 1)); 83| 45.7k| run->runs[index + 1].value = pos; 84| 45.7k| run->runs[index + 1].length = 0; 85| | return true; 86| 46.0k|} run_container_create_given_capacity: 89| 372k|run_container_t *run_container_create_given_capacity(int32_t size) { 90| 372k| run_container_t *run; 91| | /* Allocate the run container itself. */ 92| 372k| if ((run = (run_container_t *)roaring_malloc(sizeof(run_container_t))) == ------------------ | Branch (92:9): [True: 0, False: 372k] ------------------ 93| 372k| NULL) { 94| 0| return NULL; 95| 0| } 96| 372k| if (size <= 0) { // we don't want to rely on malloc(0) ------------------ | Branch (96:9): [True: 117k, False: 254k] ------------------ 97| 117k| run->runs = NULL; 98| 254k| } else if ((run->runs = (rle16_t *)roaring_malloc(sizeof(rle16_t) * ------------------ | Branch (98:16): [True: 0, False: 254k] ------------------ 99| 254k| size)) == NULL) { 100| 0| roaring_free(run); 101| 0| return NULL; 102| 0| } 103| 372k| run->capacity = size; 104| 372k| run->n_runs = 0; 105| 372k| return run; 106| 372k|} run_container_shrink_to_fit: 108| 2.54k|int run_container_shrink_to_fit(run_container_t *src) { 109| 2.54k| if (src->n_runs == src->capacity) return 0; // nothing to do ------------------ | Branch (109:9): [True: 2.54k, False: 0] ------------------ 110| 0| int savings = src->capacity - src->n_runs; 111| 0| src->capacity = src->n_runs; 112| 0| rle16_t *oldruns = src->runs; 113| 0| src->runs = 114| 0| (rle16_t *)roaring_realloc(oldruns, src->capacity * sizeof(rle16_t)); 115| 0| if (src->runs == NULL) roaring_free(oldruns); // should never happen? ------------------ | Branch (115:9): [True: 0, False: 0] ------------------ 116| 0| return savings; 117| 2.54k|} run_container_create: 119| 117k|run_container_t *run_container_create(void) { 120| 117k| return run_container_create_given_capacity(RUN_DEFAULT_INIT_SIZE); 121| 117k|} run_container_clone: 124| 164k|run_container_t *run_container_clone(const run_container_t *src) { 125| 164k| run_container_t *run = run_container_create_given_capacity(src->capacity); 126| 164k| if (run == NULL) return NULL; ------------------ | Branch (126:9): [True: 0, False: 164k] ------------------ 127| 164k| run->capacity = src->capacity; 128| 164k| run->n_runs = src->n_runs; 129| 164k| memcpy(run->runs, src->runs, src->n_runs * sizeof(rle16_t)); 130| 164k| return run; 131| 164k|} run_container_free: 193| 372k|void run_container_free(run_container_t *run) { 194| 372k| if (run == NULL) return; ------------------ | Branch (194:9): [True: 0, False: 372k] ------------------ 195| 372k| roaring_free(run->runs); 196| 372k| roaring_free(run); 197| 372k|} run_container_grow: 199| 123k|void run_container_grow(run_container_t *run, int32_t min, bool copy) { 200| 123k| int32_t newCapacity = (run->capacity == 0) ? RUN_DEFAULT_INIT_SIZE ------------------ | Branch (200:27): [True: 117k, False: 5.14k] ------------------ 201| 123k| : run->capacity < 64 ? run->capacity * 2 ------------------ | Branch (201:29): [True: 3.54k, False: 1.60k] ------------------ 202| 5.14k| : run->capacity < 1024 ? run->capacity * 3 / 2 ------------------ | Branch (202:29): [True: 1.60k, False: 0] ------------------ 203| 1.60k| : run->capacity * 5 / 4; 204| 123k| if (newCapacity < min) newCapacity = min; ------------------ | Branch (204:9): [True: 119k, False: 3.24k] ------------------ 205| 123k| run->capacity = newCapacity; 206| 123k| assert(run->capacity >= min); 207| 123k| if (copy) { ------------------ | Branch (207:9): [True: 5.14k, False: 117k] ------------------ 208| 5.14k| rle16_t *oldruns = run->runs; 209| 5.14k| run->runs = (rle16_t *)roaring_realloc(oldruns, 210| 5.14k| run->capacity * sizeof(rle16_t)); 211| 5.14k| if (run->runs == NULL) roaring_free(oldruns); ------------------ | Branch (211:13): [True: 0, False: 5.14k] ------------------ 212| 117k| } else { 213| 117k| roaring_free(run->runs); 214| 117k| run->runs = (rle16_t *)roaring_malloc(run->capacity * sizeof(rle16_t)); 215| 117k| } 216| | // We may have run->runs == NULL. 217| 123k|} run_container_copy: 220| 53.5k|void run_container_copy(const run_container_t *src, run_container_t *dst) { 221| 53.5k| const int32_t n_runs = src->n_runs; 222| 53.5k| if (src->n_runs > dst->capacity) { ------------------ | Branch (222:9): [True: 53.4k, False: 34] ------------------ 223| | run_container_grow(dst, n_runs, false); 224| 53.4k| } 225| 53.5k| dst->n_runs = n_runs; 226| 53.5k| memcpy(dst->runs, src->runs, sizeof(rle16_t) * n_runs); 227| 53.5k|} run_container_union: 232| 690| const run_container_t *src_2, run_container_t *dst) { 233| | // TODO: this could be a lot more efficient 234| | 235| | // we start out with inexpensive checks 236| 690| const bool if1 = run_container_is_full(src_1); 237| 690| const bool if2 = run_container_is_full(src_2); 238| 690| if (if1 || if2) { ------------------ | Branch (238:9): [True: 0, False: 690] | Branch (238:16): [True: 14, False: 676] ------------------ 239| 14| if (if1) { ------------------ | Branch (239:13): [True: 0, False: 14] ------------------ 240| 0| run_container_copy(src_1, dst); 241| 0| return; 242| 0| } 243| 14| if (if2) { ------------------ | Branch (243:13): [True: 14, False: 0] ------------------ 244| 14| run_container_copy(src_2, dst); 245| 14| return; 246| 14| } 247| 14| } 248| 676| const int32_t neededcapacity = src_1->n_runs + src_2->n_runs; 249| 676| if (dst->capacity < neededcapacity) ------------------ | Branch (249:9): [True: 676, False: 0] ------------------ 250| 676| run_container_grow(dst, neededcapacity, false); 251| 676| dst->n_runs = 0; 252| 676| int32_t rlepos = 0; 253| 676| int32_t xrlepos = 0; 254| | 255| 676| rle16_t previousrle; 256| 676| if (src_1->runs[rlepos].value <= src_2->runs[xrlepos].value) { ------------------ | Branch (256:9): [True: 632, False: 44] ------------------ 257| 632| previousrle = run_container_append_first(dst, src_1->runs[rlepos]); 258| 632| rlepos++; 259| 632| } else { 260| 44| previousrle = run_container_append_first(dst, src_2->runs[xrlepos]); 261| 44| xrlepos++; 262| 44| } 263| | 264| 85.1k| while ((xrlepos < src_2->n_runs) && (rlepos < src_1->n_runs)) { ------------------ | Branch (264:12): [True: 84.8k, False: 295] | Branch (264:41): [True: 84.4k, False: 381] ------------------ 265| 84.4k| rle16_t newrl; 266| 84.4k| if (src_1->runs[rlepos].value <= src_2->runs[xrlepos].value) { ------------------ | Branch (266:13): [True: 46.8k, False: 37.6k] ------------------ 267| 46.8k| newrl = src_1->runs[rlepos]; 268| 46.8k| rlepos++; 269| 46.8k| } else { 270| 37.6k| newrl = src_2->runs[xrlepos]; 271| 37.6k| xrlepos++; 272| 37.6k| } 273| 84.4k| run_container_append(dst, newrl, &previousrle); 274| 84.4k| } 275| 1.65k| while (xrlepos < src_2->n_runs) { ------------------ | Branch (275:12): [True: 975, False: 676] ------------------ 276| 975| run_container_append(dst, src_2->runs[xrlepos], &previousrle); 277| 975| xrlepos++; 278| 975| } 279| 14.7k| while (rlepos < src_1->n_runs) { ------------------ | Branch (279:12): [True: 14.0k, False: 676] ------------------ 280| 14.0k| run_container_append(dst, src_1->runs[rlepos], &previousrle); 281| 14.0k| rlepos++; 282| 14.0k| } 283| 676|} run_container_union_inplace: 288| 1.67k| const run_container_t *src_2) { 289| | // TODO: this could be a lot more efficient 290| | 291| | // we start out with inexpensive checks 292| 1.67k| const bool if1 = run_container_is_full(src_1); 293| 1.67k| const bool if2 = run_container_is_full(src_2); 294| 1.67k| if (if1 || if2) { ------------------ | Branch (294:9): [True: 0, False: 1.67k] | Branch (294:16): [True: 34, False: 1.64k] ------------------ 295| 34| if (if1) { ------------------ | Branch (295:13): [True: 0, False: 34] ------------------ 296| 0| return; 297| 0| } 298| 34| if (if2) { ------------------ | Branch (298:13): [True: 34, False: 0] ------------------ 299| 34| run_container_copy(src_2, src_1); 300| 34| return; 301| 34| } 302| 34| } 303| | // we move the data to the end of the current array 304| 1.64k| const int32_t maxoutput = src_1->n_runs + src_2->n_runs; 305| 1.64k| const int32_t neededcapacity = maxoutput + src_1->n_runs; 306| 1.64k| if (src_1->capacity < neededcapacity) ------------------ | Branch (306:9): [True: 1.64k, False: 0] ------------------ 307| 1.64k| run_container_grow(src_1, neededcapacity, true); 308| 1.64k| memmove(src_1->runs + maxoutput, src_1->runs, 309| 1.64k| src_1->n_runs * sizeof(rle16_t)); 310| 1.64k| rle16_t *inputsrc1 = src_1->runs + maxoutput; 311| 1.64k| const int32_t input1nruns = src_1->n_runs; 312| 1.64k| src_1->n_runs = 0; 313| 1.64k| int32_t rlepos = 0; 314| 1.64k| int32_t xrlepos = 0; 315| | 316| 1.64k| rle16_t previousrle; 317| 1.64k| if (inputsrc1[rlepos].value <= src_2->runs[xrlepos].value) { ------------------ | Branch (317:9): [True: 1.56k, False: 73] ------------------ 318| 1.56k| previousrle = run_container_append_first(src_1, inputsrc1[rlepos]); 319| 1.56k| rlepos++; 320| 1.56k| } else { 321| 73| previousrle = run_container_append_first(src_1, src_2->runs[xrlepos]); 322| 73| xrlepos++; 323| 73| } 324| 194k| while ((xrlepos < src_2->n_runs) && (rlepos < input1nruns)) { ------------------ | Branch (324:12): [True: 194k, False: 139] | Branch (324:41): [True: 192k, False: 1.50k] ------------------ 325| 192k| rle16_t newrl; 326| 192k| if (inputsrc1[rlepos].value <= src_2->runs[xrlepos].value) { ------------------ | Branch (326:13): [True: 93.6k, False: 99.0k] ------------------ 327| 93.6k| newrl = inputsrc1[rlepos]; 328| 93.6k| rlepos++; 329| 99.0k| } else { 330| 99.0k| newrl = src_2->runs[xrlepos]; 331| 99.0k| xrlepos++; 332| 99.0k| } 333| 192k| run_container_append(src_1, newrl, &previousrle); 334| 192k| } 335| 4.22k| while (xrlepos < src_2->n_runs) { ------------------ | Branch (335:12): [True: 2.58k, False: 1.64k] ------------------ 336| 2.58k| run_container_append(src_1, src_2->runs[xrlepos], &previousrle); 337| 2.58k| xrlepos++; 338| 2.58k| } 339| 10.4k| while (rlepos < input1nruns) { ------------------ | Branch (339:12): [True: 8.76k, False: 1.64k] ------------------ 340| 8.76k| run_container_append(src_1, inputsrc1[rlepos], &previousrle); 341| 8.76k| rlepos++; 342| 8.76k| } 343| 1.64k|} run_container_xor: 349| 55.7k| const run_container_t *src_2, run_container_t *dst) { 350| | // don't bother to convert xor with full range into negation 351| | // since negation is implemented similarly 352| | 353| 55.7k| const int32_t neededcapacity = src_1->n_runs + src_2->n_runs; 354| 55.7k| if (dst->capacity < neededcapacity) ------------------ | Branch (354:9): [True: 55.7k, False: 0] ------------------ 355| 55.7k| run_container_grow(dst, neededcapacity, false); 356| | 357| 55.7k| int32_t pos1 = 0; 358| 55.7k| int32_t pos2 = 0; 359| 55.7k| dst->n_runs = 0; 360| | 361| 235k| while ((pos1 < src_1->n_runs) && (pos2 < src_2->n_runs)) { ------------------ | Branch (361:12): [True: 179k, False: 55.4k] | Branch (361:38): [True: 179k, False: 315] ------------------ 362| 179k| if (src_1->runs[pos1].value <= src_2->runs[pos2].value) { ------------------ | Branch (362:13): [True: 120k, False: 58.8k] ------------------ 363| 120k| run_container_smart_append_exclusive(dst, src_1->runs[pos1].value, 364| 120k| src_1->runs[pos1].length); 365| 120k| pos1++; 366| 120k| } else { 367| 58.8k| run_container_smart_append_exclusive(dst, src_2->runs[pos2].value, 368| 58.8k| src_2->runs[pos2].length); 369| 58.8k| pos2++; 370| 58.8k| } 371| 179k| } 372| 71.1k| while (pos1 < src_1->n_runs) { ------------------ | Branch (372:12): [True: 15.4k, False: 55.7k] ------------------ 373| 15.4k| run_container_smart_append_exclusive(dst, src_1->runs[pos1].value, 374| 15.4k| src_1->runs[pos1].length); 375| 15.4k| pos1++; 376| 15.4k| } 377| | 378| 113k| while (pos2 < src_2->n_runs) { ------------------ | Branch (378:12): [True: 57.8k, False: 55.7k] ------------------ 379| 57.8k| run_container_smart_append_exclusive(dst, src_2->runs[pos2].value, 380| 57.8k| src_2->runs[pos2].length); 381| 57.8k| pos2++; 382| 57.8k| } 383| 55.7k|} run_container_intersection: 389| 56.7k| run_container_t *dst) { 390| 56.7k| const bool if1 = run_container_is_full(src_1); 391| 56.7k| const bool if2 = run_container_is_full(src_2); 392| 56.7k| if (if1 || if2) { ------------------ | Branch (392:9): [True: 53.3k, False: 3.39k] | Branch (392:16): [True: 14, False: 3.38k] ------------------ 393| 53.3k| if (if1) { ------------------ | Branch (393:13): [True: 53.3k, False: 14] ------------------ 394| 53.3k| run_container_copy(src_2, dst); 395| 53.3k| return; 396| 53.3k| } 397| 14| if (if2) { ------------------ | Branch (397:13): [True: 14, False: 0] ------------------ 398| 14| run_container_copy(src_1, dst); 399| 14| return; 400| 14| } 401| 14| } 402| | // TODO: this could be a lot more efficient, could use SIMD optimizations 403| 3.38k| const int32_t neededcapacity = src_1->n_runs + src_2->n_runs; 404| 3.38k| if (dst->capacity < neededcapacity) ------------------ | Branch (404:9): [True: 3.38k, False: 0] ------------------ 405| 3.38k| run_container_grow(dst, neededcapacity, false); 406| 3.38k| dst->n_runs = 0; 407| 3.38k| int32_t rlepos = 0; 408| 3.38k| int32_t xrlepos = 0; 409| 3.38k| int32_t start = src_1->runs[rlepos].value; 410| 3.38k| int32_t end = start + src_1->runs[rlepos].length + 1; 411| 3.38k| int32_t xstart = src_2->runs[xrlepos].value; 412| 3.38k| int32_t xend = xstart + src_2->runs[xrlepos].length + 1; 413| 187k| while ((rlepos < src_1->n_runs) && (xrlepos < src_2->n_runs)) { ------------------ | Branch (413:12): [True: 184k, False: 2.92k] | Branch (413:40): [True: 183k, False: 460] ------------------ 414| 183k| if (end <= xstart) { ------------------ | Branch (414:13): [True: 48.0k, False: 135k] ------------------ 415| 48.0k| ++rlepos; 416| 48.0k| if (rlepos < src_1->n_runs) { ------------------ | Branch (416:17): [True: 47.9k, False: 73] ------------------ 417| 47.9k| start = src_1->runs[rlepos].value; 418| 47.9k| end = start + src_1->runs[rlepos].length + 1; 419| 47.9k| } 420| 135k| } else if (xend <= start) { ------------------ | Branch (420:20): [True: 12.0k, False: 123k] ------------------ 421| 12.0k| ++xrlepos; 422| 12.0k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (422:17): [True: 12.0k, False: 42] ------------------ 423| 12.0k| xstart = src_2->runs[xrlepos].value; 424| 12.0k| xend = xstart + src_2->runs[xrlepos].length + 1; 425| 12.0k| } 426| 123k| } else { // they overlap 427| 123k| const int32_t lateststart = start > xstart ? start : xstart; ------------------ | Branch (427:41): [True: 17.8k, False: 106k] ------------------ 428| 123k| int32_t earliestend; 429| 123k| if (end == xend) { // improbable ------------------ | Branch (429:17): [True: 91.3k, False: 32.4k] ------------------ 430| 91.3k| earliestend = end; 431| 91.3k| rlepos++; 432| 91.3k| xrlepos++; 433| 91.3k| if (rlepos < src_1->n_runs) { ------------------ | Branch (433:21): [True: 88.7k, False: 2.69k] ------------------ 434| 88.7k| start = src_1->runs[rlepos].value; 435| 88.7k| end = start + src_1->runs[rlepos].length + 1; 436| 88.7k| } 437| 91.3k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (437:21): [True: 88.5k, False: 2.89k] ------------------ 438| 88.5k| xstart = src_2->runs[xrlepos].value; 439| 88.5k| xend = xstart + src_2->runs[xrlepos].length + 1; 440| 88.5k| } 441| 91.3k| } else if (end < xend) { ------------------ | Branch (441:24): [True: 18.1k, False: 14.3k] ------------------ 442| 18.1k| earliestend = end; 443| 18.1k| rlepos++; 444| 18.1k| if (rlepos < src_1->n_runs) { ------------------ | Branch (444:21): [True: 18.0k, False: 160] ------------------ 445| 18.0k| start = src_1->runs[rlepos].value; 446| 18.0k| end = start + src_1->runs[rlepos].length + 1; 447| 18.0k| } 448| | 449| 18.1k| } else { // end > xend 450| 14.3k| earliestend = xend; 451| 14.3k| xrlepos++; 452| 14.3k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (452:21): [True: 14.1k, False: 134] ------------------ 453| 14.1k| xstart = src_2->runs[xrlepos].value; 454| 14.1k| xend = xstart + src_2->runs[xrlepos].length + 1; 455| 14.1k| } 456| 14.3k| } 457| 123k| dst->runs[dst->n_runs].value = (uint16_t)lateststart; 458| 123k| dst->runs[dst->n_runs].length = 459| 123k| (uint16_t)(earliestend - lateststart - 1); 460| 123k| dst->n_runs++; 461| 123k| } 462| 183k| } 463| 3.38k|} run_container_intersection_cardinality: 467| 2.76k| const run_container_t *src_2) { 468| 2.76k| const bool if1 = run_container_is_full(src_1); 469| 2.76k| const bool if2 = run_container_is_full(src_2); 470| 2.76k| if (if1 || if2) { ------------------ | Branch (470:9): [True: 0, False: 2.76k] | Branch (470:16): [True: 56, False: 2.70k] ------------------ 471| 56| if (if1) { ------------------ | Branch (471:13): [True: 0, False: 56] ------------------ 472| 0| return run_container_cardinality(src_2); 473| 0| } 474| 56| if (if2) { ------------------ | Branch (474:13): [True: 56, False: 0] ------------------ 475| 56| return run_container_cardinality(src_1); 476| 56| } 477| 56| } 478| 2.70k| int answer = 0; 479| 2.70k| int32_t rlepos = 0; 480| 2.70k| int32_t xrlepos = 0; 481| 2.70k| int32_t start = src_1->runs[rlepos].value; 482| 2.70k| int32_t end = start + src_1->runs[rlepos].length + 1; 483| 2.70k| int32_t xstart = src_2->runs[xrlepos].value; 484| 2.70k| int32_t xend = xstart + src_2->runs[xrlepos].length + 1; 485| 297k| while ((rlepos < src_1->n_runs) && (xrlepos < src_2->n_runs)) { ------------------ | Branch (485:12): [True: 295k, False: 1.88k] | Branch (485:40): [True: 294k, False: 820] ------------------ 486| 294k| if (end <= xstart) { ------------------ | Branch (486:13): [True: 70.3k, False: 224k] ------------------ 487| 70.3k| ++rlepos; 488| 70.3k| if (rlepos < src_1->n_runs) { ------------------ | Branch (488:17): [True: 70.0k, False: 292] ------------------ 489| 70.0k| start = src_1->runs[rlepos].value; 490| 70.0k| end = start + src_1->runs[rlepos].length + 1; 491| 70.0k| } 492| 224k| } else if (xend <= start) { ------------------ | Branch (492:20): [True: 48.3k, False: 175k] ------------------ 493| 48.3k| ++xrlepos; 494| 48.3k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (494:17): [True: 48.1k, False: 168] ------------------ 495| 48.1k| xstart = src_2->runs[xrlepos].value; 496| 48.1k| xend = xstart + src_2->runs[xrlepos].length + 1; 497| 48.1k| } 498| 175k| } else { // they overlap 499| 175k| const int32_t lateststart = start > xstart ? start : xstart; ------------------ | Branch (499:41): [True: 71.2k, False: 104k] ------------------ 500| 175k| int32_t earliestend; 501| 175k| if (end == xend) { // improbable ------------------ | Branch (501:17): [True: 84.5k, False: 91.2k] ------------------ 502| 84.5k| earliestend = end; 503| 84.5k| rlepos++; 504| 84.5k| xrlepos++; 505| 84.5k| if (rlepos < src_1->n_runs) { ------------------ | Branch (505:21): [True: 83.5k, False: 952] ------------------ 506| 83.5k| start = src_1->runs[rlepos].value; 507| 83.5k| end = start + src_1->runs[rlepos].length + 1; 508| 83.5k| } 509| 84.5k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (509:21): [True: 83.4k, False: 1.08k] ------------------ 510| 83.4k| xstart = src_2->runs[xrlepos].value; 511| 83.4k| xend = xstart + src_2->runs[xrlepos].length + 1; 512| 83.4k| } 513| 91.2k| } else if (end < xend) { ------------------ | Branch (513:24): [True: 72.6k, False: 18.6k] ------------------ 514| 72.6k| earliestend = end; 515| 72.6k| rlepos++; 516| 72.6k| if (rlepos < src_1->n_runs) { ------------------ | Branch (516:21): [True: 72.0k, False: 640] ------------------ 517| 72.0k| start = src_1->runs[rlepos].value; 518| 72.0k| end = start + src_1->runs[rlepos].length + 1; 519| 72.0k| } 520| | 521| 72.6k| } else { // end > xend 522| 18.6k| earliestend = xend; 523| 18.6k| xrlepos++; 524| 18.6k| if (xrlepos < src_2->n_runs) { ------------------ | Branch (524:21): [True: 18.4k, False: 200] ------------------ 525| 18.4k| xstart = src_2->runs[xrlepos].value; 526| 18.4k| xend = xstart + src_2->runs[xrlepos].length + 1; 527| 18.4k| } 528| 18.6k| } 529| 175k| answer += earliestend - lateststart; 530| 175k| } 531| 294k| } 532| 2.70k| return answer; 533| 2.76k|} run_container_intersect: 536| 690| const run_container_t *src_2) { 537| 690| const bool if1 = run_container_is_full(src_1); 538| 690| const bool if2 = run_container_is_full(src_2); 539| 690| if (if1 || if2) { ------------------ | Branch (539:9): [True: 0, False: 690] | Branch (539:16): [True: 14, False: 676] ------------------ 540| 14| if (if1) { ------------------ | Branch (540:13): [True: 0, False: 14] ------------------ 541| 0| return !run_container_empty(src_2); 542| 0| } 543| 14| if (if2) { ------------------ | Branch (543:13): [True: 14, False: 0] ------------------ 544| 14| return !run_container_empty(src_1); 545| 14| } 546| 14| } 547| 676| int32_t rlepos = 0; 548| 676| int32_t xrlepos = 0; 549| 676| int32_t start = src_1->runs[rlepos].value; 550| 676| int32_t end = start + src_1->runs[rlepos].length + 1; 551| 676| int32_t xstart = src_2->runs[xrlepos].value; 552| 676| int32_t xend = xstart + src_2->runs[xrlepos].length + 1; 553| 3.04k| while ((rlepos < src_1->n_runs) && (xrlepos < src_2->n_runs)) { ------------------ | Branch (553:12): [True: 3.02k, False: 13] | Branch (553:40): [True: 3.02k, False: 2] ------------------ 554| 3.02k| if (end <= xstart) { ------------------ | Branch (554:13): [True: 1.97k, False: 1.05k] ------------------ 555| 1.97k| ++rlepos; 556| 1.97k| if (rlepos < src_1->n_runs) { ------------------ | Branch (556:17): [True: 1.96k, False: 13] ------------------ 557| 1.96k| start = src_1->runs[rlepos].value; 558| 1.96k| end = start + src_1->runs[rlepos].length + 1; 559| 1.96k| } 560| 1.97k| } else if (xend <= start) { ------------------ | Branch (560:20): [True: 393, False: 661] ------------------ 561| 393| ++xrlepos; 562| 393| if (xrlepos < src_2->n_runs) { ------------------ | Branch (562:17): [True: 391, False: 2] ------------------ 563| 391| xstart = src_2->runs[xrlepos].value; 564| 391| xend = xstart + src_2->runs[xrlepos].length + 1; 565| 391| } 566| 661| } else { // they overlap 567| 661| return true; 568| 661| } 569| 3.02k| } 570| 15| return false; 571| 676|} run_container_andnot: 576| 1.13k| const run_container_t *src_2, run_container_t *dst) { 577| | // following Java implementation as of June 2016 578| | 579| 1.13k| if (dst->capacity < src_1->n_runs + src_2->n_runs) ------------------ | Branch (579:9): [True: 1.13k, False: 0] ------------------ 580| 1.13k| run_container_grow(dst, src_1->n_runs + src_2->n_runs, false); 581| | 582| 1.13k| dst->n_runs = 0; 583| | 584| 1.13k| int rlepos1 = 0; 585| 1.13k| int rlepos2 = 0; 586| 1.13k| int32_t start = src_1->runs[rlepos1].value; 587| 1.13k| int32_t end = start + src_1->runs[rlepos1].length + 1; 588| 1.13k| int32_t start2 = src_2->runs[rlepos2].value; 589| 1.13k| int32_t end2 = start2 + src_2->runs[rlepos2].length + 1; 590| | 591| 168k| while ((rlepos1 < src_1->n_runs) && (rlepos2 < src_2->n_runs)) { ------------------ | Branch (591:12): [True: 168k, False: 692] | Branch (591:41): [True: 167k, False: 447] ------------------ 592| 167k| if (end <= start2) { ------------------ | Branch (592:13): [True: 24.2k, False: 143k] ------------------ 593| | // output the first run 594| 24.2k| dst->runs[dst->n_runs++] = 595| 24.2k| CROARING_MAKE_RLE16(start, end - start - 1); ------------------ | | 60| 24.2k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 596| 24.2k| rlepos1++; 597| 24.2k| if (rlepos1 < src_1->n_runs) { ------------------ | Branch (597:17): [True: 24.1k, False: 73] ------------------ 598| 24.1k| start = src_1->runs[rlepos1].value; 599| 24.1k| end = start + src_1->runs[rlepos1].length + 1; 600| 24.1k| } 601| 143k| } else if (end2 <= start) { ------------------ | Branch (601:20): [True: 68.4k, False: 75.0k] ------------------ 602| | // exit the second run 603| 68.4k| rlepos2++; 604| 68.4k| if (rlepos2 < src_2->n_runs) { ------------------ | Branch (604:17): [True: 68.0k, False: 447] ------------------ 605| 68.0k| start2 = src_2->runs[rlepos2].value; 606| 68.0k| end2 = start2 + src_2->runs[rlepos2].length + 1; 607| 68.0k| } 608| 75.0k| } else { 609| 75.0k| if (start < start2) { ------------------ | Branch (609:17): [True: 21.4k, False: 53.5k] ------------------ 610| 21.4k| dst->runs[dst->n_runs++] = 611| 21.4k| CROARING_MAKE_RLE16(start, start2 - start - 1); ------------------ | | 60| 21.4k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 612| 21.4k| } 613| 75.0k| if (end2 < end) { ------------------ | Branch (613:17): [True: 21.0k, False: 53.9k] ------------------ 614| 21.0k| start = end2; 615| 53.9k| } else { 616| 53.9k| rlepos1++; 617| 53.9k| if (rlepos1 < src_1->n_runs) { ------------------ | Branch (617:21): [True: 53.3k, False: 619] ------------------ 618| 53.3k| start = src_1->runs[rlepos1].value; 619| 53.3k| end = start + src_1->runs[rlepos1].length + 1; 620| 53.3k| } 621| 53.9k| } 622| 75.0k| } 623| 167k| } 624| 1.13k| if (rlepos1 < src_1->n_runs) { ------------------ | Branch (624:9): [True: 447, False: 692] ------------------ 625| 447| dst->runs[dst->n_runs++] = CROARING_MAKE_RLE16(start, end - start - 1); ------------------ | | 60| 447| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 626| 447| rlepos1++; 627| 447| if (rlepos1 < src_1->n_runs) { ------------------ | Branch (627:13): [True: 210, False: 237] ------------------ 628| 210| memcpy(dst->runs + dst->n_runs, src_1->runs + rlepos1, 629| 210| sizeof(rle16_t) * (src_1->n_runs - rlepos1)); 630| 210| dst->n_runs += src_1->n_runs - rlepos1; 631| 210| } 632| 447| } 633| 1.13k|} run_container_read: 739| 49| const char *buf) { 740| 49| (void)cardinality; 741| 49| uint16_t cast_16; 742| 49| memcpy(&cast_16, buf, sizeof(uint16_t)); 743| 49| container->n_runs = croaring_letoh16(cast_16); ------------------ | | 523| 49|#define croaring_letoh16(x) croaring_htole16(x) | | ------------------ | | | | 517| 49|#define croaring_htole16(x) (x) | | ------------------ ------------------ 744| 49| if (container->n_runs > container->capacity) ------------------ | Branch (744:9): [True: 7, False: 42] ------------------ 745| 7| run_container_grow(container, container->n_runs, false); 746| 49| if (container->n_runs > 0) { ------------------ | Branch (746:9): [True: 7, False: 42] ------------------ 747| |#if CROARING_IS_BIG_ENDIAN 748| | const char *in = buf + sizeof(uint16_t); 749| | for (int32_t i = 0; i < container->n_runs; ++i) { 750| | uint16_t v_le, l_le; 751| | memcpy(&v_le, in, sizeof(uint16_t)); 752| | memcpy(&l_le, in + sizeof(uint16_t), sizeof(uint16_t)); 753| | container->runs[i].value = croaring_letoh16(v_le); 754| | container->runs[i].length = croaring_letoh16(l_le); 755| | in += sizeof(rle16_t); 756| | } 757| |#else 758| 7| memcpy(container->runs, buf + sizeof(uint16_t), 759| 7| container->n_runs * sizeof(rle16_t)); 760| 7|#endif 761| 7| } 762| 49| return run_container_size_in_bytes(container); 763| 49|} run_container_iterate: 766| 55.6k| roaring_iterator iterator, void *ptr) { 767| 156k| for (int i = 0; i < cont->n_runs; ++i) { ------------------ | Branch (767:21): [True: 100k, False: 55.6k] ------------------ 768| 100k| uint32_t run_start = base + cont->runs[i].value; 769| 100k| uint16_t le = cont->runs[i].length; 770| | 771| 3.56G| for (int j = 0; j <= le; ++j) ------------------ | Branch (771:25): [True: 3.56G, False: 100k] ------------------ 772| 3.56G| if (!iterator(run_start + j, ptr)) return false; ------------------ | Branch (772:17): [True: 0, False: 3.56G] ------------------ 773| 100k| } 774| 55.6k| return true; 775| 55.6k|} run_container_is_subset: 792| 256| const run_container_t *container2) { 793| 256| int i1 = 0, i2 = 0; 794| 856| while (i1 < container1->n_runs && i2 < container2->n_runs) { ------------------ | Branch (794:12): [True: 856, False: 0] | Branch (794:39): [True: 834, False: 22] ------------------ 795| 834| int start1 = container1->runs[i1].value; 796| 834| int stop1 = start1 + container1->runs[i1].length; 797| 834| int start2 = container2->runs[i2].value; 798| 834| int stop2 = start2 + container2->runs[i2].length; 799| 834| if (start1 < start2) { ------------------ | Branch (799:13): [True: 234, False: 600] ------------------ 800| 234| return false; 801| 600| } else { // start1 >= start2 802| 600| if (stop1 < stop2) { ------------------ | Branch (802:17): [True: 0, False: 600] ------------------ 803| 0| i1++; 804| 600| } else if (stop1 == stop2) { ------------------ | Branch (804:24): [True: 0, False: 600] ------------------ 805| 0| i1++; 806| 0| i2++; 807| 600| } else { // stop1 > stop2 808| 600| i2++; 809| 600| } 810| 600| } 811| 834| } 812| 22| if (i1 == container1->n_runs) { ------------------ | Branch (812:9): [True: 0, False: 22] ------------------ 813| 0| return true; 814| 22| } else { 815| | return false; 816| 22| } 817| 22|} run_container_smart_append_exclusive: 826| 390k| const uint16_t length) { 827| 390k| int old_end; 828| 390k| rle16_t *last_run = src->n_runs ? src->runs + (src->n_runs - 1) : NULL; ------------------ | Branch (828:25): [True: 319k, False: 70.7k] ------------------ 829| 390k| rle16_t *appended_last_run = src->runs + src->n_runs; 830| | 831| 390k| if (!src->n_runs || ------------------ | Branch (831:9): [True: 70.7k, False: 319k] ------------------ 832| 319k| (start > (old_end = last_run->value + last_run->length + 1))) { ------------------ | Branch (832:9): [True: 175k, False: 144k] ------------------ 833| 245k| *appended_last_run = CROARING_MAKE_RLE16(start, length); ------------------ | | 60| 245k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 834| 245k| src->n_runs++; 835| 245k| return; 836| 245k| } 837| 144k| if (old_end == start) { ------------------ | Branch (837:9): [True: 3.50k, False: 140k] ------------------ 838| | // we merge 839| 3.50k| last_run->length += (length + 1); 840| 3.50k| return; 841| 3.50k| } 842| 140k| int new_end = start + length + 1; 843| | 844| 140k| if (start == last_run->value) { ------------------ | Branch (844:9): [True: 109k, False: 31.0k] ------------------ 845| | // wipe out previous 846| 109k| if (new_end < old_end) { ------------------ | Branch (846:13): [True: 6.83k, False: 103k] ------------------ 847| 6.83k| *last_run = CROARING_MAKE_RLE16(new_end, old_end - new_end - 1); ------------------ | | 60| 6.83k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 848| 6.83k| return; 849| 103k| } else if (new_end > old_end) { ------------------ | Branch (849:20): [True: 2.32k, False: 100k] ------------------ 850| 2.32k| *last_run = CROARING_MAKE_RLE16(old_end, new_end - old_end - 1); ------------------ | | 60| 2.32k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 851| 2.32k| return; 852| 100k| } else { 853| 100k| src->n_runs--; 854| 100k| return; 855| 100k| } 856| 109k| } 857| 31.0k| last_run->length = start - last_run->value - 1; 858| 31.0k| if (new_end < old_end) { ------------------ | Branch (858:9): [True: 25.1k, False: 5.84k] ------------------ 859| 25.1k| *appended_last_run = 860| 25.1k| CROARING_MAKE_RLE16(new_end, old_end - new_end - 1); ------------------ | | 60| 25.1k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 861| 25.1k| src->n_runs++; 862| 25.1k| } else if (new_end > old_end) { ------------------ | Branch (862:16): [True: 1.05k, False: 4.79k] ------------------ 863| 1.05k| *appended_last_run = 864| 1.05k| CROARING_MAKE_RLE16(old_end, new_end - old_end - 1); ------------------ | | 60| 1.05k| (rle16_t) { .value = (uint16_t)(val), .length = (uint16_t)(len) } ------------------ 865| 1.05k| src->n_runs++; 866| 1.05k| } 867| 31.0k|} run_container_select: 871| 2.54k| uint32_t *element) { 872| 13.1k| for (int i = 0; i < container->n_runs; i++) { ------------------ | Branch (872:21): [True: 13.1k, False: 89] ------------------ 873| 13.1k| uint16_t length = container->runs[i].length; 874| 13.1k| if (rank <= *start_rank + length) { ------------------ | Branch (874:13): [True: 2.45k, False: 10.6k] ------------------ 875| 2.45k| uint16_t value = container->runs[i].value; 876| 2.45k| *element = value + rank - (*start_rank); 877| 2.45k| return true; 878| 2.45k| } else 879| 10.6k| *start_rank += length + 1; 880| 13.1k| } 881| 89| return false; 882| 2.54k|} run_container_rank: 884| 2.54k|int run_container_rank(const run_container_t *container, uint16_t x) { 885| 2.54k| int sum = 0; 886| 2.54k| uint32_t x32 = x; 887| 10.3k| for (int i = 0; i < container->n_runs; i++) { ------------------ | Branch (887:21): [True: 10.2k, False: 53] ------------------ 888| 10.2k| uint32_t startpoint = container->runs[i].value; 889| 10.2k| uint32_t length = container->runs[i].length; 890| 10.2k| uint32_t endpoint = length + startpoint; 891| 10.2k| if (x <= endpoint) { ------------------ | Branch (891:13): [True: 2.49k, False: 7.75k] ------------------ 892| 2.49k| if (x < startpoint) break; ------------------ | Branch (892:17): [True: 253, False: 2.24k] ------------------ 893| 2.24k| return sum + (x32 - startpoint) + 1; 894| 7.75k| } else { 895| 7.75k| sum += length + 1; 896| 7.75k| } 897| 10.2k| } 898| 306| return sum; 899| 2.54k|} _avx2_run_container_to_uint32_array: 1021| 56.2k| uint32_t base) { 1022| 56.2k| int outpos = 0; 1023| 56.2k| uint32_t *out = (uint32_t *)vout; 1024| | 1025| 197k| for (int i = 0; i < cont->n_runs; ++i) { ------------------ | Branch (1025:21): [True: 140k, False: 56.2k] ------------------ 1026| 140k| uint32_t run_start = base + cont->runs[i].value; 1027| 140k| uint16_t le = cont->runs[i].length; 1028| 140k| if (le < 8) { ------------------ | Branch (1028:13): [True: 80.8k, False: 59.9k] ------------------ 1029| 208k| for (int j = 0; j <= le; ++j) { ------------------ | Branch (1029:29): [True: 127k, False: 80.8k] ------------------ 1030| 127k| uint32_t val = run_start + j; 1031| 127k| memcpy(out + outpos, &val, 1032| 127k| sizeof(uint32_t)); // should be compiled as a MOV on x64 1033| 127k| outpos++; 1034| 127k| } 1035| 80.8k| } else { 1036| 59.9k| int j = 0; 1037| 59.9k| __m256i run_start_v = _mm256_set1_epi32(run_start); 1038| | // [8,8,8,8....] 1039| 59.9k| __m256i inc = _mm256_set1_epi32(8); 1040| | // used for generate sequence: 1041| | // [0, 1, 2, 3...], [8, 9, 10,...] 1042| 59.9k| __m256i delta = _mm256_setr_epi32(0, 1, 2, 3, 4, 5, 6, 7); 1043| 446M| for (j = 0; j + 8 <= le; j += 8) { ------------------ | Branch (1043:25): [True: 446M, False: 59.9k] ------------------ 1044| 446M| __m256i val_v = _mm256_add_epi32(run_start_v, delta); 1045| 446M| _mm256_storeu_si256((__m256i *)(out + outpos), val_v); 1046| 446M| delta = _mm256_add_epi32(inc, delta); 1047| 446M| outpos += 8; 1048| 446M| } 1049| 515k| for (; j <= le; ++j) { ------------------ | Branch (1049:20): [True: 455k, False: 59.9k] ------------------ 1050| 455k| uint32_t val = run_start + j; 1051| 455k| memcpy(out + outpos, &val, 1052| 455k| sizeof(uint32_t)); // should be compiled as a MOV on x64 1053| 455k| outpos++; 1054| 455k| } 1055| 59.9k| } 1056| 140k| } 1057| 56.2k| return outpos; 1058| 56.2k|} run_container_cardinality: 1077| 504k|int run_container_cardinality(const run_container_t *run) { 1078| | // Empirically AVX-512 is not always faster than AVX2 1079| |#if CROARING_COMPILER_SUPPORTS_AVX512 && \ 1080| | CROARING_ENABLE_AVX512_RUN_CONTAINER_CARDINALITY 1081| | if (croaring_hardware_support() & ROARING_SUPPORTS_AVX512) { 1082| | return _avx512_run_container_cardinality(run); 1083| | } else 1084| |#endif 1085| 504k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (1085:13): [True: 504k, False: 0] ------------------ 1086| 504k| return _avx2_run_container_cardinality(run); 1087| 504k| } else { 1088| 0| return _scalar_run_container_cardinality(run); 1089| 0| } 1090| 504k|} run_container_to_uint32_array: 1111| 56.2k| uint32_t base) { 1112| 56.2k| if (croaring_hardware_support() & ROARING_SUPPORTS_AVX2) { ------------------ | Branch (1112:9): [True: 56.2k, False: 0] ------------------ 1113| 56.2k| return _avx2_run_container_to_uint32_array(vout, cont, base); 1114| 56.2k| } else { 1115| 0| return _scalar_run_container_to_uint32_array(vout, cont, base); 1116| 0| } 1117| 56.2k|} run.c:_avx2_run_container_cardinality: 991| 504k|static inline int _avx2_run_container_cardinality(const run_container_t *run) { 992| 504k| const int32_t n_runs = run->n_runs; 993| 504k| const rle16_t *runs = run->runs; 994| | 995| | /* by initializing with n_runs, we omit counting the +1 for each pair. */ 996| 504k| int sum = n_runs; 997| 504k| int32_t k = 0; 998| 504k| const int32_t step = sizeof(__m256i) / sizeof(rle16_t); 999| 504k| if (n_runs > step) { ------------------ | Branch (999:9): [True: 33.4k, False: 470k] ------------------ 1000| 33.4k| __m256i total = _mm256_setzero_si256(); 1001| 377k| for (; k + step <= n_runs; k += step) { ------------------ | Branch (1001:16): [True: 343k, False: 33.4k] ------------------ 1002| 343k| __m256i ymm1 = _mm256_lddqu_si256((const __m256i *)(runs + k)); 1003| 343k| __m256i justlengths = _mm256_srli_epi32(ymm1, 16); 1004| 343k| total = _mm256_add_epi32(total, justlengths); 1005| 343k| } 1006| | // a store might be faster than extract? 1007| 33.4k| uint32_t buffer[sizeof(__m256i) / sizeof(rle16_t)]; 1008| 33.4k| _mm256_storeu_si256((__m256i *)buffer, total); 1009| 33.4k| sum += (buffer[0] + buffer[1]) + (buffer[2] + buffer[3]) + 1010| 33.4k| (buffer[4] + buffer[5]) + (buffer[6] + buffer[7]); 1011| 33.4k| } 1012| 1.05M| for (; k < n_runs; ++k) { ------------------ | Branch (1012:12): [True: 552k, False: 504k] ------------------ 1013| 552k| sum += runs[k].length; 1014| 552k| } 1015| | 1016| 504k| return sum; 1017| 504k|} croaring_hardware_support: 321| 894k|int croaring_hardware_support(void) { 322| 894k| static 323| 894k|#if CROARING_ATOMIC_IMPL == CROARING_ATOMIC_IMPL_C 324| 894k| _Atomic 325| 894k|#endif 326| 894k| int support = 0xFFFFFFF; 327| 894k| if (support == 0xFFFFFFF) { ------------------ | Branch (327:9): [True: 1, False: 894k] ------------------ 328| 1| bool has_avx2 = (croaring_detect_supported_architectures() & 329| 1| CROARING_AVX2) == CROARING_AVX2; 330| 1| bool has_avx512 = false; 331| 1|#if CROARING_COMPILER_SUPPORTS_AVX512 332| 1| has_avx512 = (croaring_detect_supported_architectures() & 333| 1| CROARING_AVX512_REQUIRED) == CROARING_AVX512_REQUIRED; 334| 1|#endif // CROARING_COMPILER_SUPPORTS_AVX512 335| 1| support = (has_avx2 ? ROARING_SUPPORTS_AVX2 : 0) | ------------------ | Branch (335:20): [True: 1, False: 0] ------------------ 336| 1| (has_avx512 ? ROARING_SUPPORTS_AVX512 : 0); ------------------ | Branch (336:20): [True: 0, False: 1] ------------------ 337| 1| } 338| 894k| return support; 339| 894k|} isadetection.c:croaring_detect_supported_architectures: 269| 2|static uint32_t croaring_detect_supported_architectures(void) { 270| | // we use an atomic for thread safety 271| 2| static _Atomic uint32_t buffer = CROARING_UNINITIALIZED; 272| 2| if (buffer == CROARING_UNINITIALIZED) { ------------------ | Branch (272:9): [True: 1, False: 1] ------------------ 273| | // atomicity is sufficient 274| 1| buffer = dynamic_croaring_detect_supported_architectures(); 275| 1| } 276| 2| return buffer; 277| 2|} isadetection.c:dynamic_croaring_detect_supported_architectures: 154| 1|static inline uint32_t dynamic_croaring_detect_supported_architectures(void) { 155| 1| uint32_t eax, ebx, ecx, edx; 156| 1| uint32_t host_isa = 0x0; 157| | // Can be found on Intel ISA Reference for CPUID 158| 1| static uint32_t cpuid_avx2_bit = 159| 1| 1 << 5; ///< @private Bit 5 of EBX for EAX=0x7 160| 1| static uint32_t cpuid_bmi1_bit = 161| 1| 1 << 3; ///< @private bit 3 of EBX for EAX=0x7 162| 1| static uint32_t cpuid_bmi2_bit = 163| 1| 1 << 8; ///< @private bit 8 of EBX for EAX=0x7 164| 1| static uint32_t cpuid_avx512f_bit = 165| 1| 1 << 16; ///< @private bit 16 of EBX for EAX=0x7 166| 1| static uint32_t cpuid_avx512dq_bit = 167| 1| 1 << 17; ///< @private bit 17 of EBX for EAX=0x7 168| 1| static uint32_t cpuid_avx512bw_bit = 169| 1| 1 << 30; ///< @private bit 30 of EBX for EAX=0x7 170| 1| static uint32_t cpuid_avx512vbmi2_bit = 171| 1| 1 << 6; ///< @private bit 6 of ECX for EAX=0x7 172| 1| static uint32_t cpuid_avx512bitalg_bit = 173| 1| 1 << 12; ///< @private bit 12 of ECX for EAX=0x7 174| 1| static uint32_t cpuid_avx512vpopcntdq_bit = 175| 1| 1 << 14; ///< @private bit 14 of ECX for EAX=0x7 176| 1| static uint64_t cpuid_avx256_saved = 1 << 2; ///< @private bit 2 = AVX 177| 1| static uint64_t cpuid_avx512_saved = 178| 1| 7 << 5; ///< @private bits 5,6,7 = opmask, ZMM_hi256, hi16_ZMM 179| 1| static uint32_t cpuid_sse42_bit = 180| 1| 1 << 20; ///< @private bit 20 of ECX for EAX=0x1 181| 1| static uint32_t cpuid_osxsave = 182| 1| (1 << 26) | (1 << 27); ///< @private bits 26+27 of ECX for EAX=0x1 183| 1| static uint32_t cpuid_pclmulqdq_bit = 184| 1| 1 << 1; ///< @private bit 1 of ECX for EAX=0x1 185| | 186| | // EBX for EAX=0x1 187| 1| eax = 0x1; 188| 1| ecx = 0x0; 189| 1| cpuid(&eax, &ebx, &ecx, &edx); 190| | 191| 1| if (ecx & cpuid_sse42_bit) { ------------------ | Branch (191:9): [True: 1, False: 0] ------------------ 192| 1| host_isa |= CROARING_SSE42; 193| 1| } else { 194| 0| return host_isa; // everything after is redundant 195| 0| } 196| | 197| 1| if (ecx & cpuid_pclmulqdq_bit) { ------------------ | Branch (197:9): [True: 1, False: 0] ------------------ 198| 1| host_isa |= CROARING_PCLMULQDQ; 199| 1| } 200| | 201| 1| if ((ecx & cpuid_osxsave) != cpuid_osxsave) { ------------------ | Branch (201:9): [True: 0, False: 1] ------------------ 202| 0| return host_isa; 203| 0| } 204| | 205| | // xgetbv for checking if the OS saves registers 206| 1| uint64_t xcr0 = xgetbv(); 207| | 208| 1| if ((xcr0 & cpuid_avx256_saved) == 0) { ------------------ | Branch (208:9): [True: 0, False: 1] ------------------ 209| 0| return host_isa; 210| 0| } 211| | 212| | // ECX for EAX=0x7 213| 1| eax = 0x7; 214| 1| ecx = 0x0; 215| 1| cpuid(&eax, &ebx, &ecx, &edx); 216| 1| if (ebx & cpuid_avx2_bit) { ------------------ | Branch (216:9): [True: 1, False: 0] ------------------ 217| 1| host_isa |= CROARING_AVX2; 218| 1| } 219| 1| if (ebx & cpuid_bmi1_bit) { ------------------ | Branch (219:9): [True: 1, False: 0] ------------------ 220| 1| host_isa |= CROARING_BMI1; 221| 1| } 222| | 223| 1| if (ebx & cpuid_bmi2_bit) { ------------------ | Branch (223:9): [True: 1, False: 0] ------------------ 224| 1| host_isa |= CROARING_BMI2; 225| 1| } 226| | 227| 1| if (!((xcr0 & cpuid_avx512_saved) == cpuid_avx512_saved)) { ------------------ | Branch (227:9): [True: 1, False: 0] ------------------ 228| 1| return host_isa; 229| 1| } 230| | 231| 0| if (ebx & cpuid_avx512f_bit) { ------------------ | Branch (231:9): [True: 0, False: 0] ------------------ 232| 0| host_isa |= CROARING_AVX512F; 233| 0| } 234| | 235| 0| if (ebx & cpuid_avx512bw_bit) { ------------------ | Branch (235:9): [True: 0, False: 0] ------------------ 236| 0| host_isa |= CROARING_AVX512BW; 237| 0| } 238| | 239| 0| if (ebx & cpuid_avx512dq_bit) { ------------------ | Branch (239:9): [True: 0, False: 0] ------------------ 240| 0| host_isa |= CROARING_AVX512DQ; 241| 0| } 242| | 243| 0| if (ecx & cpuid_avx512vbmi2_bit) { ------------------ | Branch (243:9): [True: 0, False: 0] ------------------ 244| 0| host_isa |= CROARING_AVX512VBMI2; 245| 0| } 246| | 247| 0| if (ecx & cpuid_avx512bitalg_bit) { ------------------ | Branch (247:9): [True: 0, False: 0] ------------------ 248| 0| host_isa |= CROARING_AVX512BITALG; 249| 0| } 250| | 251| 0| if (ecx & cpuid_avx512vpopcntdq_bit) { ------------------ | Branch (251:9): [True: 0, False: 0] ------------------ 252| 0| host_isa |= CROARING_AVX512VPOPCNTDQ; 253| 0| } 254| | 255| 0| return host_isa; 256| 1|} isadetection.c:cpuid: 115| 2| uint32_t *edx) { 116| |#if CROARING_REGULAR_VISUAL_STUDIO 117| | int cpu_info[4]; 118| | __cpuidex(cpu_info, *eax, *ecx); 119| | *eax = cpu_info[0]; 120| | *ebx = cpu_info[1]; 121| | *ecx = cpu_info[2]; 122| | *edx = cpu_info[3]; 123| |#elif (defined(HAVE_GCC_GET_CPUID) && defined(USE_GCC_GET_CPUID)) || \ 124| | defined(__FILC__) 125| | uint32_t level = *eax; 126| | __get_cpuid(level, eax, ebx, ecx, edx); 127| |#else 128| 2| uint32_t a = *eax, b, c = *ecx, d; 129| 2| __asm__("cpuid\n\t" : "+a"(a), "=b"(b), "+c"(c), "=d"(d)); 130| 2| *eax = a; 131| 2| *ebx = b; 132| 2| *ecx = c; 133| 2| *edx = d; 134| 2|#endif 135| 2|} isadetection.c:xgetbv: 137| 1|static inline uint64_t xgetbv(void) { 138| |#if defined(_MSC_VER) 139| | return _xgetbv(0); 140| |#elif defined(__FILC__) 141| | return zxgetbv(); 142| |#else 143| 1| uint32_t xcr0_lo, xcr0_hi; 144| 1| __asm__("xgetbv\n\t" : "=a"(xcr0_lo), "=d"(xcr0_hi) : "c"(0)); 145| 1| return xcr0_lo | ((uint64_t)xcr0_hi << 32); 146| 1|#endif 147| 1|} roaring_malloc: 48| 1.33M|void* roaring_malloc(size_t n) { return global_memory_hook.malloc(n); } roaring_realloc: 50| 115k|void* roaring_realloc(void* p, size_t new_sz) { 51| 115k| return global_memory_hook.realloc(p, new_sz); 52| 115k|} roaring_free: 58| 1.56M|void roaring_free(void* p) { global_memory_hook.free(p); } roaring_aligned_malloc: 60| 273k|void* roaring_aligned_malloc(size_t alignment, size_t size) { 61| 273k| return global_memory_hook.aligned_malloc(alignment, size); 62| 273k|} roaring_aligned_free: 64| 273k|void roaring_aligned_free(void* p) { global_memory_hook.aligned_free(p); } memory.c:roaring_bitmap_aligned_malloc: 11| 273k|static void* roaring_bitmap_aligned_malloc(size_t alignment, size_t size) { 12| 273k| void* p; 13| |#ifdef _MSC_VER 14| | p = _aligned_malloc(size, alignment); 15| |#elif defined(__MINGW32__) || defined(__MINGW64__) 16| | p = __mingw_aligned_malloc(size, alignment); 17| |#else 18| | // somehow, if this is used before including "x86intrin.h", it creates an 19| | // implicit defined warning. 20| 273k| if (posix_memalign(&p, alignment, size) != 0) return NULL; ------------------ | Branch (20:9): [True: 0, False: 273k] ------------------ 21| 273k|#endif 22| 273k| return p; 23| 273k|} memory.c:roaring_bitmap_aligned_free: 25| 273k|static void roaring_bitmap_aligned_free(void* memblock) { 26| |#ifdef _MSC_VER 27| | _aligned_free(memblock); 28| |#elif defined(__MINGW32__) || defined(__MINGW64__) 29| | __mingw_aligned_free(memblock); 30| |#else 31| 273k| free(memblock); 32| 273k|#endif 33| 273k|} roaring_bitmap_create_with_capacity: 86| 24.8k|roaring_bitmap_t *roaring_bitmap_create_with_capacity(uint32_t cap) { 87| 24.8k| roaring_bitmap_t *ans = 88| 24.8k| (roaring_bitmap_t *)roaring_malloc(sizeof(roaring_bitmap_t)); 89| 24.8k| if (!ans) { ------------------ | Branch (89:9): [True: 0, False: 24.8k] ------------------ 90| 0| return NULL; 91| 0| } 92| 24.8k| bool is_ok = ra_init_with_capacity(&ans->high_low_container, cap); 93| 24.8k| if (!is_ok) { ------------------ | Branch (93:9): [True: 0, False: 24.8k] ------------------ 94| 0| roaring_free(ans); 95| 0| return NULL; 96| 0| } 97| 24.8k| return ans; 98| 24.8k|} roaring_bitmap_init_with_capacity: 100| 31.2k|bool roaring_bitmap_init_with_capacity(roaring_bitmap_t *r, uint32_t cap) { 101| 31.2k| return ra_init_with_capacity(&r->high_low_container, cap); 102| 31.2k|} roaring_bitmap_add_many: 135| 18.7k| const uint32_t *vals) { 136| 18.7k| uint32_t val; 137| 18.7k| const uint32_t *start = vals; 138| 18.7k| const uint32_t *end = vals + n_args; 139| 18.7k| const uint32_t *current_val = start; 140| | 141| 18.7k| if (n_args == 0) { ------------------ | Branch (141:9): [True: 0, False: 18.7k] ------------------ 142| 0| return; 143| 0| } 144| | 145| 18.7k| uint8_t typecode; 146| 18.7k| int idx; 147| 18.7k| container_t *container; 148| 18.7k| val = *current_val; 149| 18.7k| container = containerptr_roaring_bitmap_add(r, val, &typecode, &idx); 150| 18.7k| roaring_bulk_context_t context = {container, idx, (uint16_t)(val >> 16), 151| 18.7k| typecode}; 152| | 153| 7.82M| for (; current_val != end; current_val++) { ------------------ | Branch (153:12): [True: 7.80M, False: 18.7k] ------------------ 154| 7.80M| memcpy(&val, current_val, sizeof(val)); 155| 7.80M| add_bulk_impl(r, &context, val); 156| 7.80M| } 157| 18.7k|} roaring_bitmap_add_range_closed: 252| 2.71k| uint32_t max) { 253| 2.71k| if (min > max) { ------------------ | Branch (253:9): [True: 0, False: 2.71k] ------------------ 254| 0| return; 255| 0| } 256| | 257| 2.71k| roaring_array_t *ra = &r->high_low_container; 258| | 259| 2.71k| uint32_t min_key = min >> 16; 260| 2.71k| uint32_t max_key = max >> 16; 261| | 262| 2.71k| int32_t num_required_containers = max_key - min_key + 1; 263| 2.71k| int32_t suffix_length = 264| 2.71k| count_greater(ra->keys, ra->size, (uint16_t)max_key); 265| 2.71k| int32_t prefix_length = 266| 2.71k| count_less(ra->keys, ra->size - suffix_length, (uint16_t)min_key); 267| 2.71k| int32_t common_length = ra->size - prefix_length - suffix_length; 268| | 269| 2.71k| if (num_required_containers > common_length) { ------------------ | Branch (269:9): [True: 1.24k, False: 1.46k] ------------------ 270| 1.24k| ra_shift_tail(ra, suffix_length, 271| 1.24k| num_required_containers - common_length); 272| 1.24k| } 273| | 274| 2.71k| int32_t src = prefix_length + common_length - 1; 275| 2.71k| int32_t dst = ra->size - suffix_length - 1; 276| 74.1k| for (uint32_t key = max_key; key != min_key - 1; ------------------ | Branch (276:34): [True: 71.4k, False: 2.71k] ------------------ 277| 71.4k| key--) { // beware of min_key==0 278| 71.4k| uint32_t container_min = (min_key == key) ? (min & 0xffff) : 0; ------------------ | Branch (278:34): [True: 2.71k, False: 68.7k] ------------------ 279| 71.4k| uint32_t container_max = (max_key == key) ? (max & 0xffff) : 0xffff; ------------------ | Branch (279:34): [True: 2.71k, False: 68.7k] ------------------ 280| 71.4k| container_t *new_container; 281| 71.4k| uint8_t new_type; 282| | 283| 71.4k| if (src >= 0 && ra->keys[src] == key) { ------------------ | Branch (283:13): [True: 71.4k, False: 0] | Branch (283:25): [True: 3.16k, False: 68.2k] ------------------ 284| 3.16k| ra_unshare_container_at_index(ra, (uint16_t)src); 285| 3.16k| new_container = 286| 3.16k| container_add_range(ra->containers[src], ra->typecodes[src], 287| 3.16k| container_min, container_max, &new_type); 288| 3.16k| if (new_container != ra->containers[src]) { ------------------ | Branch (288:17): [True: 2.10k, False: 1.05k] ------------------ 289| 2.10k| container_free(ra->containers[src], ra->typecodes[src]); 290| 2.10k| } 291| 3.16k| src--; 292| 68.2k| } else { 293| 68.2k| new_container = container_from_range(&new_type, container_min, 294| 68.2k| container_max + 1, 1); 295| 68.2k| } 296| 71.4k| ra_replace_key_and_container_at_index(ra, dst, (uint16_t)key, 297| 71.4k| new_container, new_type); 298| 71.4k| dst--; 299| 71.4k| } 300| 2.71k|} roaring_bitmap_remove_range_closed: 303| 7.14k| uint32_t max) { 304| 7.14k| if (min > max) { ------------------ | Branch (304:9): [True: 767, False: 6.37k] ------------------ 305| 767| return; 306| 767| } 307| | 308| 6.37k| roaring_array_t *ra = &r->high_low_container; 309| | 310| 6.37k| uint32_t min_key = min >> 16; 311| 6.37k| uint32_t max_key = max >> 16; 312| | 313| 6.37k| int32_t src = count_less(ra->keys, ra->size, (uint16_t)min_key); 314| 6.37k| int32_t dst = src; 315| 26.8k| while (src < ra->size && ra->keys[src] <= max_key) { ------------------ | Branch (315:12): [True: 22.4k, False: 4.37k] | Branch (315:30): [True: 20.4k, False: 2.00k] ------------------ 316| 20.4k| uint32_t container_min = 317| 20.4k| (min_key == ra->keys[src]) ? (min & 0xffff) : 0; ------------------ | Branch (317:13): [True: 5.99k, False: 14.4k] ------------------ 318| 20.4k| uint32_t container_max = 319| 20.4k| (max_key == ra->keys[src]) ? (max & 0xffff) : 0xffff; ------------------ | Branch (319:13): [True: 5.65k, False: 14.7k] ------------------ 320| 20.4k| ra_unshare_container_at_index(ra, (uint16_t)src); 321| 20.4k| container_t *new_container; 322| 20.4k| uint8_t new_type; 323| 20.4k| new_container = 324| 20.4k| container_remove_range(ra->containers[src], ra->typecodes[src], 325| 20.4k| container_min, container_max, &new_type); 326| 20.4k| if (new_container != ra->containers[src]) { ------------------ | Branch (326:13): [True: 14.5k, False: 5.85k] ------------------ 327| 14.5k| container_free(ra->containers[src], ra->typecodes[src]); 328| 14.5k| } 329| 20.4k| if (new_container) { ------------------ | Branch (329:13): [True: 6.11k, False: 14.3k] ------------------ 330| 6.11k| ra_replace_key_and_container_at_index(ra, dst, ra->keys[src], 331| 6.11k| new_container, new_type); 332| 6.11k| dst++; 333| 6.11k| } 334| 20.4k| src++; 335| 20.4k| } 336| 6.37k| if (src > dst) { ------------------ | Branch (336:9): [True: 615, False: 5.76k] ------------------ 337| 615| ra_shift_tail(ra, ra->size - src, dst - src); 338| 615| } 339| 6.37k|} roaring_bitmap_copy: 525| 153|roaring_bitmap_t *roaring_bitmap_copy(const roaring_bitmap_t *r) { 526| 153| roaring_bitmap_t *ans = 527| 153| (roaring_bitmap_t *)roaring_malloc(sizeof(roaring_bitmap_t)); 528| 153| if (!ans) { ------------------ | Branch (528:9): [True: 0, False: 153] ------------------ 529| 0| return NULL; 530| 0| } 531| 153| if (!ra_init_with_capacity( // allocation of list of containers can fail ------------------ | Branch (531:9): [True: 0, False: 153] ------------------ 532| 153| &ans->high_low_container, r->high_low_container.size)) { 533| 0| roaring_free(ans); 534| 0| return NULL; 535| 0| } 536| 153| if (!ra_overwrite( // memory allocation of individual containers may fail ------------------ | Branch (536:9): [True: 0, False: 153] ------------------ 537| 153| &r->high_low_container, &ans->high_low_container, is_cow(r))) { 538| 0| roaring_bitmap_free(ans); // overwrite should leave in freeable state 539| 0| return NULL; 540| 0| } 541| 153| roaring_bitmap_set_copy_on_write(ans, is_cow(r)); 542| 153| return ans; 543| 153|} roaring_bitmap_overwrite: 546| 13.9k| const roaring_bitmap_t *src) { 547| 13.9k| roaring_bitmap_set_copy_on_write(dest, is_cow(src)); 548| 13.9k| return ra_overwrite(&src->high_low_container, &dest->high_low_container, 549| 13.9k| is_cow(src)); 550| 13.9k|} roaring_bitmap_clear: 562| 62.4k|void roaring_bitmap_clear(roaring_bitmap_t *r) { 563| 62.4k| ra_reset(&r->high_low_container); 564| 62.4k|} roaring_bitmap_add: 566| 6.24k|void roaring_bitmap_add(roaring_bitmap_t *r, uint32_t val) { 567| 6.24k| roaring_array_t *ra = &r->high_low_container; 568| | 569| 6.24k| const uint16_t hb = val >> 16; 570| 6.24k| const int i = ra_get_index(ra, hb); 571| 6.24k| uint8_t typecode; 572| 6.24k| if (i >= 0) { ------------------ | Branch (572:9): [True: 4.54k, False: 1.70k] ------------------ 573| 4.54k| ra_unshare_container_at_index(ra, (uint16_t)i); 574| 4.54k| container_t *container = 575| 4.54k| ra_get_container_at_index(ra, (uint16_t)i, &typecode); 576| 4.54k| uint8_t newtypecode = typecode; 577| 4.54k| container_t *container2 = 578| 4.54k| container_add(container, val & 0xFFFF, typecode, &newtypecode); 579| 4.54k| if (container2 != container) { ------------------ | Branch (579:13): [True: 0, False: 4.54k] ------------------ 580| 0| container_free(container, typecode); 581| 0| ra_set_container_at_index(&r->high_low_container, i, container2, 582| 0| newtypecode); 583| 0| } 584| 4.54k| } else { 585| 1.70k| array_container_t *newac = array_container_create(); 586| 1.70k| container_t *container = 587| 1.70k| container_add(newac, val & 0xFFFF, ARRAY_CONTAINER_TYPE, &typecode); ------------------ | | 49| 1.70k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 588| | // we could just assume that it stays an array container 589| 1.70k| ra_insert_new_key_value_at(&r->high_low_container, -i - 1, hb, 590| 1.70k| container, typecode); 591| 1.70k| } 592| 6.24k|} roaring_bitmap_add_checked: 594| 6.24k|bool roaring_bitmap_add_checked(roaring_bitmap_t *r, uint32_t val) { 595| 6.24k| const uint16_t hb = val >> 16; 596| 6.24k| const int i = ra_get_index(&r->high_low_container, hb); 597| 6.24k| uint8_t typecode; 598| 6.24k| bool result = false; 599| 6.24k| if (i >= 0) { ------------------ | Branch (599:9): [True: 4.78k, False: 1.45k] ------------------ 600| 4.78k| ra_unshare_container_at_index(&r->high_low_container, (uint16_t)i); 601| 4.78k| container_t *container = ra_get_container_at_index( 602| 4.78k| &r->high_low_container, (uint16_t)i, &typecode); 603| | 604| 4.78k| const int oldCardinality = 605| 4.78k| container_get_cardinality(container, typecode); 606| | 607| 4.78k| uint8_t newtypecode = typecode; 608| 4.78k| container_t *container2 = 609| 4.78k| container_add(container, val & 0xFFFF, typecode, &newtypecode); 610| 4.78k| if (container2 != container) { ------------------ | Branch (610:13): [True: 0, False: 4.78k] ------------------ 611| 0| container_free(container, typecode); 612| 0| ra_set_container_at_index(&r->high_low_container, i, container2, 613| 0| newtypecode); 614| 0| result = true; 615| 4.78k| } else { 616| 4.78k| const int newCardinality = 617| 4.78k| container_get_cardinality(container, newtypecode); 618| | 619| 4.78k| result = oldCardinality != newCardinality; 620| 4.78k| } 621| 4.78k| } else { 622| 1.45k| array_container_t *newac = array_container_create(); 623| 1.45k| container_t *container = 624| 1.45k| container_add(newac, val & 0xFFFF, ARRAY_CONTAINER_TYPE, &typecode); ------------------ | | 49| 1.45k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 625| | // we could just assume that it stays an array container 626| 1.45k| ra_insert_new_key_value_at(&r->high_low_container, -i - 1, hb, 627| 1.45k| container, typecode); 628| 1.45k| result = true; 629| 1.45k| } 630| | 631| 6.24k| return result; 632| 6.24k|} roaring_bitmap_remove: 634| 6.24k|void roaring_bitmap_remove(roaring_bitmap_t *r, uint32_t val) { 635| 6.24k| const uint16_t hb = val >> 16; 636| 6.24k| const int i = ra_get_index(&r->high_low_container, hb); 637| 6.24k| uint8_t typecode; 638| 6.24k| if (i >= 0) { ------------------ | Branch (638:9): [True: 5.67k, False: 573] ------------------ 639| 5.67k| ra_unshare_container_at_index(&r->high_low_container, (uint16_t)i); 640| 5.67k| container_t *container = ra_get_container_at_index( 641| 5.67k| &r->high_low_container, (uint16_t)i, &typecode); 642| 5.67k| uint8_t newtypecode = typecode; 643| 5.67k| container_t *container2 = 644| 5.67k| container_remove(container, val & 0xFFFF, typecode, &newtypecode); 645| 5.67k| if (container2 != container) { ------------------ | Branch (645:13): [True: 15, False: 5.65k] ------------------ 646| 15| container_free(container, typecode); 647| 15| ra_set_container_at_index(&r->high_low_container, i, container2, 648| 15| newtypecode); 649| 15| } 650| 5.67k| if (container_nonzero_cardinality(container2, newtypecode)) { ------------------ | Branch (650:13): [True: 5.62k, False: 49] ------------------ 651| 5.62k| ra_set_container_at_index(&r->high_low_container, i, container2, 652| 5.62k| newtypecode); 653| 5.62k| } else { 654| 49| ra_remove_at_index_and_free(&r->high_low_container, i); 655| 49| } 656| 5.67k| } 657| 6.24k|} roaring_bitmap_remove_checked: 659| 6.24k|bool roaring_bitmap_remove_checked(roaring_bitmap_t *r, uint32_t val) { 660| 6.24k| const uint16_t hb = val >> 16; 661| 6.24k| const int i = ra_get_index(&r->high_low_container, hb); 662| 6.24k| uint8_t typecode; 663| 6.24k| bool result = false; 664| 6.24k| if (i >= 0) { ------------------ | Branch (664:9): [True: 5.67k, False: 568] ------------------ 665| 5.67k| ra_unshare_container_at_index(&r->high_low_container, (uint16_t)i); 666| 5.67k| container_t *container = ra_get_container_at_index( 667| 5.67k| &r->high_low_container, (uint16_t)i, &typecode); 668| | 669| 5.67k| const int oldCardinality = 670| 5.67k| container_get_cardinality(container, typecode); 671| | 672| 5.67k| uint8_t newtypecode = typecode; 673| 5.67k| container_t *container2 = 674| 5.67k| container_remove(container, val & 0xFFFF, typecode, &newtypecode); 675| 5.67k| if (container2 != container) { ------------------ | Branch (675:13): [True: 3, False: 5.67k] ------------------ 676| 3| container_free(container, typecode); 677| 3| ra_set_container_at_index(&r->high_low_container, i, container2, 678| 3| newtypecode); 679| 3| } 680| | 681| 5.67k| const int newCardinality = 682| 5.67k| container_get_cardinality(container2, newtypecode); 683| | 684| 5.67k| if (newCardinality != 0) { ------------------ | Branch (684:13): [True: 5.66k, False: 10] ------------------ 685| 5.66k| ra_set_container_at_index(&r->high_low_container, i, container2, 686| 5.66k| newtypecode); 687| 5.66k| } else { 688| 10| ra_remove_at_index_and_free(&r->high_low_container, i); 689| 10| } 690| | 691| 5.67k| result = oldCardinality != newCardinality; 692| 5.67k| } 693| 6.24k| return result; 694| 6.24k|} roaring_bitmap_and: 732| 6.24k| const roaring_bitmap_t *x2) { 733| 6.24k| uint8_t result_type = 0; 734| 6.24k| const int length1 = x1->high_low_container.size, 735| 6.24k| length2 = x2->high_low_container.size; 736| 6.24k| uint32_t neededcap = length1 > length2 ? length2 : length1; ------------------ | Branch (736:26): [True: 51, False: 6.19k] ------------------ 737| 6.24k| roaring_bitmap_t *answer = roaring_bitmap_create_with_capacity(neededcap); 738| 6.24k| roaring_bitmap_set_copy_on_write(answer, is_cow(x1) || is_cow(x2)); ------------------ | Branch (738:46): [True: 0, False: 6.24k] | Branch (738:60): [True: 0, False: 6.24k] ------------------ 739| | 740| 6.24k| int pos1 = 0, pos2 = 0; 741| | 742| 12.4k| while (pos1 < length1 && pos2 < length2) { ------------------ | Branch (742:12): [True: 6.24k, False: 6.19k] | Branch (742:30): [True: 6.19k, False: 51] ------------------ 743| 6.19k| const uint16_t s1 = 744| 6.19k| ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 745| 6.19k| const uint16_t s2 = 746| 6.19k| ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 747| | 748| 6.19k| if (s1 == s2) { ------------------ | Branch (748:13): [True: 6.12k, False: 67] ------------------ 749| 6.12k| uint8_t type1, type2; 750| 6.12k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 751| 6.12k| (uint16_t)pos1, &type1); 752| 6.12k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 753| 6.12k| (uint16_t)pos2, &type2); 754| 6.12k| container_t *c = container_and(c1, type1, c2, type2, &result_type); 755| | 756| 6.12k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (756:17): [True: 5.81k, False: 316] ------------------ 757| 5.81k| ra_append(&answer->high_low_container, s1, c, result_type); 758| 5.81k| } else { 759| 316| container_free(c, result_type); // otherwise: memory leak! 760| 316| } 761| 6.12k| ++pos1; 762| 6.12k| ++pos2; 763| 6.12k| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (763:20): [True: 67, False: 0] ------------------ 764| 67| pos1 = ra_advance_until(&x1->high_low_container, s2, pos1); 765| 67| } else { // s1 > s2 766| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 767| 0| } 768| 6.19k| } 769| 6.24k| return answer; 770| 6.24k|} roaring_bitmap_and_inplace: 813| 6.24k| const roaring_bitmap_t *x2) { 814| 6.24k| if (x1 == x2) return; ------------------ | Branch (814:9): [True: 0, False: 6.24k] ------------------ 815| 6.24k| int pos1 = 0, pos2 = 0, intersection_size = 0; 816| 6.24k| const int length1 = ra_get_size(&x1->high_low_container); 817| 6.24k| const int length2 = ra_get_size(&x2->high_low_container); 818| | 819| | // any skipped-over or newly emptied containers in x1 820| | // have to be freed. 821| 69.6k| while (pos1 < length1 && pos2 < length2) { ------------------ | Branch (821:12): [True: 63.4k, False: 6.19k] | Branch (821:30): [True: 63.3k, False: 51] ------------------ 822| 63.3k| const uint16_t s1 = 823| 63.3k| ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 824| 63.3k| const uint16_t s2 = 825| 63.3k| ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 826| | 827| 63.3k| if (s1 == s2) { ------------------ | Branch (827:13): [True: 63.2k, False: 67] ------------------ 828| 63.2k| uint8_t type1, type2, result_type; 829| 63.2k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 830| 63.2k| (uint16_t)pos1, &type1); 831| 63.2k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 832| 63.2k| (uint16_t)pos2, &type2); 833| | 834| | // We do the computation "in place" only when c1 is not a shared 835| | // container. Rationale: using a shared container safely with in 836| | // place computation would require making a copy and then doing the 837| | // computation in place which is likely less efficient than avoiding 838| | // in place entirely and always generating a new container. 839| 63.2k| container_t *c = 840| 63.2k| (type1 == SHARED_CONTAINER_TYPE) ------------------ | | 51| 63.2k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (840:17): [True: 0, False: 63.2k] ------------------ 841| 63.2k| ? container_and(c1, type1, c2, type2, &result_type) 842| 63.2k| : container_iand(c1, type1, c2, type2, &result_type); 843| | 844| 63.2k| if (c != c1) { // in this instance a new container was created, and ------------------ | Branch (844:17): [True: 57.3k, False: 5.97k] ------------------ 845| | // we need to free the old one 846| 57.3k| container_free(c1, type1); 847| 57.3k| } 848| 63.2k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (848:17): [True: 63.2k, False: 0] ------------------ 849| 63.2k| ra_replace_key_and_container_at_index(&x1->high_low_container, 850| 63.2k| intersection_size, s1, c, 851| 63.2k| result_type); 852| 63.2k| intersection_size++; 853| 63.2k| } else { 854| 0| container_free(c, result_type); 855| 0| } 856| 63.2k| ++pos1; 857| 63.2k| ++pos2; 858| 63.2k| } else if (s1 < s2) { ------------------ | Branch (858:20): [True: 67, False: 0] ------------------ 859| 67| pos1 = ra_advance_until_freeing(&x1->high_low_container, s2, pos1); 860| 67| } else { // s1 > s2 861| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 862| 0| } 863| 63.3k| } 864| | 865| | // if we ended early because x2 ran out, then all remaining in x1 should be 866| | // freed 867| 6.29k| while (pos1 < length1) { ------------------ | Branch (867:12): [True: 51, False: 6.24k] ------------------ 868| 51| container_free(x1->high_low_container.containers[pos1], 869| 51| x1->high_low_container.typecodes[pos1]); 870| 51| ++pos1; 871| 51| } 872| | 873| | // all containers after this have either been copied or freed 874| 6.24k| ra_downsize(&x1->high_low_container, intersection_size); 875| 6.24k|} roaring_bitmap_or: 878| 6.24k| const roaring_bitmap_t *x2) { 879| 6.24k| uint8_t result_type = 0; 880| 6.24k| const int length1 = x1->high_low_container.size, 881| 6.24k| length2 = x2->high_low_container.size; 882| 6.24k| if (0 == length1) { ------------------ | Branch (882:9): [True: 0, False: 6.24k] ------------------ 883| 0| return roaring_bitmap_copy(x2); 884| 0| } 885| 6.24k| if (0 == length2) { ------------------ | Branch (885:9): [True: 51, False: 6.19k] ------------------ 886| 51| return roaring_bitmap_copy(x1); 887| 51| } 888| 6.19k| roaring_bitmap_t *answer = 889| 6.19k| roaring_bitmap_create_with_capacity(length1 + length2); 890| 6.19k| roaring_bitmap_set_copy_on_write(answer, is_cow(x1) || is_cow(x2)); ------------------ | Branch (890:46): [True: 0, False: 6.19k] | Branch (890:60): [True: 0, False: 6.19k] ------------------ 891| 6.19k| int pos1 = 0, pos2 = 0; 892| 6.19k| uint8_t type1, type2; 893| 6.19k| uint16_t s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 894| 6.19k| uint16_t s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 895| 6.19k| while (true) { ------------------ | Branch (895:12): [True: 6.19k, Folded] ------------------ 896| 6.19k| if (s1 == s2) { ------------------ | Branch (896:13): [True: 6.12k, False: 67] ------------------ 897| 6.12k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 898| 6.12k| (uint16_t)pos1, &type1); 899| 6.12k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 900| 6.12k| (uint16_t)pos2, &type2); 901| 6.12k| container_t *c = container_or(c1, type1, c2, type2, &result_type); 902| | 903| | // since we assume that the initial containers are non-empty, the 904| | // result here 905| | // can only be non-empty 906| 6.12k| ra_append(&answer->high_low_container, s1, c, result_type); 907| 6.12k| ++pos1; 908| 6.12k| ++pos2; 909| 6.12k| if (pos1 == length1) break; ------------------ | Branch (909:17): [True: 6.12k, False: 0] ------------------ 910| 0| if (pos2 == length2) break; ------------------ | Branch (910:17): [True: 0, False: 0] ------------------ 911| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 912| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 913| | 914| 67| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (914:20): [True: 67, False: 0] ------------------ 915| 67| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 916| 67| (uint16_t)pos1, &type1); 917| | // c1 = container_clone(c1, type1); 918| 67| c1 = get_copy_of_container(c1, &type1, is_cow(x1)); 919| 67| if (is_cow(x1)) { ------------------ | Branch (919:17): [True: 0, False: 67] ------------------ 920| 0| ra_set_container_at_index(&x1->high_low_container, pos1, c1, 921| 0| type1); 922| 0| } 923| 67| ra_append(&answer->high_low_container, s1, c1, type1); 924| 67| pos1++; 925| 67| if (pos1 == length1) break; ------------------ | Branch (925:17): [True: 67, False: 0] ------------------ 926| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 927| | 928| 0| } else { // s1 > s2 929| 0| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 930| 0| (uint16_t)pos2, &type2); 931| | // c2 = container_clone(c2, type2); 932| 0| c2 = get_copy_of_container(c2, &type2, is_cow(x2)); 933| 0| if (is_cow(x2)) { ------------------ | Branch (933:17): [True: 0, False: 0] ------------------ 934| 0| ra_set_container_at_index(&x2->high_low_container, pos2, c2, 935| 0| type2); 936| 0| } 937| 0| ra_append(&answer->high_low_container, s2, c2, type2); 938| 0| pos2++; 939| 0| if (pos2 == length2) break; ------------------ | Branch (939:17): [True: 0, False: 0] ------------------ 940| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 941| 0| } 942| 6.19k| } 943| 6.19k| if (pos1 == length1) { ------------------ | Branch (943:9): [True: 6.19k, False: 0] ------------------ 944| 6.19k| ra_append_copy_range(&answer->high_low_container, 945| 6.19k| &x2->high_low_container, pos2, length2, 946| 6.19k| is_cow(x2)); 947| 6.19k| } else if (pos2 == length2) { ------------------ | Branch (947:16): [True: 0, False: 0] ------------------ 948| 0| ra_append_copy_range(&answer->high_low_container, 949| 0| &x1->high_low_container, pos1, length1, 950| 0| is_cow(x1)); 951| 0| } 952| 6.19k| return answer; 953| 6.24k|} roaring_bitmap_or_inplace: 1064| 6.24k| const roaring_bitmap_t *x2) { 1065| 6.24k| uint8_t result_type = 0; 1066| 6.24k| int length1 = x1->high_low_container.size; 1067| 6.24k| const int length2 = x2->high_low_container.size; 1068| | 1069| 6.24k| if (0 == length2) return; ------------------ | Branch (1069:9): [True: 0, False: 6.24k] ------------------ 1070| | 1071| 6.24k| if (0 == length1) { ------------------ | Branch (1071:9): [True: 0, False: 6.24k] ------------------ 1072| 0| roaring_bitmap_overwrite(x1, x2); 1073| 0| return; 1074| 0| } 1075| 6.24k| int pos1 = 0, pos2 = 0; 1076| 6.24k| uint8_t type1, type2; 1077| 6.24k| uint16_t s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1078| 6.24k| uint16_t s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1079| 6.24k| while (true) { ------------------ | Branch (1079:12): [True: 6.24k, Folded] ------------------ 1080| 6.24k| if (s1 == s2) { ------------------ | Branch (1080:13): [True: 6.24k, False: 0] ------------------ 1081| 6.24k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1082| 6.24k| (uint16_t)pos1, &type1); 1083| 6.24k| if (!container_is_full(c1, type1)) { ------------------ | Branch (1083:17): [True: 6.24k, False: 0] ------------------ 1084| 6.24k| container_t *c2 = ra_get_container_at_index( 1085| 6.24k| &x2->high_low_container, (uint16_t)pos2, &type2); 1086| 6.24k| container_t *c = 1087| 6.24k| (type1 == SHARED_CONTAINER_TYPE) ------------------ | | 51| 6.24k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1087:21): [True: 0, False: 6.24k] ------------------ 1088| 6.24k| ? container_or(c1, type1, c2, type2, &result_type) 1089| 6.24k| : container_ior(c1, type1, c2, type2, &result_type); 1090| | 1091| 6.24k| if (c != c1) { // in this instance a new container was created, ------------------ | Branch (1091:21): [True: 4.49k, False: 1.75k] ------------------ 1092| | // and we need to free the old one 1093| 4.49k| container_free(c1, type1); 1094| 4.49k| } 1095| 6.24k| ra_set_container_at_index(&x1->high_low_container, pos1, c, 1096| 6.24k| result_type); 1097| 6.24k| } 1098| 6.24k| ++pos1; 1099| 6.24k| ++pos2; 1100| 6.24k| if (pos1 == length1) break; ------------------ | Branch (1100:17): [True: 6.24k, False: 0] ------------------ 1101| 0| if (pos2 == length2) break; ------------------ | Branch (1101:17): [True: 0, False: 0] ------------------ 1102| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1103| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1104| | 1105| 0| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (1105:20): [True: 0, False: 0] ------------------ 1106| 0| pos1++; 1107| 0| if (pos1 == length1) break; ------------------ | Branch (1107:17): [True: 0, False: 0] ------------------ 1108| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1109| | 1110| 0| } else { // s1 > s2 1111| 0| roaring_inplace_merge_bulk(x1, x2, pos1, pos1, pos2, false); 1112| 0| return; 1113| 0| } 1114| 6.24k| } 1115| 6.24k| if (pos1 == length1) { ------------------ | Branch (1115:9): [True: 6.24k, False: 0] ------------------ 1116| 6.24k| ra_append_copy_range(&x1->high_low_container, &x2->high_low_container, 1117| 6.24k| pos2, length2, is_cow(x2)); 1118| 6.24k| } 1119| 6.24k|} roaring_bitmap_xor: 1122| 6.24k| const roaring_bitmap_t *x2) { 1123| 6.24k| uint8_t result_type = 0; 1124| 6.24k| const int length1 = x1->high_low_container.size, 1125| 6.24k| length2 = x2->high_low_container.size; 1126| 6.24k| if (0 == length1) { ------------------ | Branch (1126:9): [True: 0, False: 6.24k] ------------------ 1127| 0| return roaring_bitmap_copy(x2); 1128| 0| } 1129| 6.24k| if (0 == length2) { ------------------ | Branch (1129:9): [True: 51, False: 6.19k] ------------------ 1130| 51| return roaring_bitmap_copy(x1); 1131| 51| } 1132| 6.19k| roaring_bitmap_t *answer = 1133| 6.19k| roaring_bitmap_create_with_capacity(length1 + length2); 1134| 6.19k| roaring_bitmap_set_copy_on_write(answer, is_cow(x1) || is_cow(x2)); ------------------ | Branch (1134:46): [True: 0, False: 6.19k] | Branch (1134:60): [True: 0, False: 6.19k] ------------------ 1135| 6.19k| int pos1 = 0, pos2 = 0; 1136| 6.19k| uint8_t type1, type2; 1137| 6.19k| uint16_t s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1138| 6.19k| uint16_t s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1139| 6.19k| while (true) { ------------------ | Branch (1139:12): [True: 6.19k, Folded] ------------------ 1140| 6.19k| if (s1 == s2) { ------------------ | Branch (1140:13): [True: 6.12k, False: 67] ------------------ 1141| 6.12k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1142| 6.12k| (uint16_t)pos1, &type1); 1143| 6.12k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 1144| 6.12k| (uint16_t)pos2, &type2); 1145| 6.12k| container_t *c = container_xor(c1, type1, c2, type2, &result_type); 1146| | 1147| 6.12k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (1147:17): [True: 6.11k, False: 17] ------------------ 1148| 6.11k| ra_append(&answer->high_low_container, s1, c, result_type); 1149| 6.11k| } else { 1150| 17| container_free(c, result_type); 1151| 17| } 1152| 6.12k| ++pos1; 1153| 6.12k| ++pos2; 1154| 6.12k| if (pos1 == length1) break; ------------------ | Branch (1154:17): [True: 6.12k, False: 0] ------------------ 1155| 0| if (pos2 == length2) break; ------------------ | Branch (1155:17): [True: 0, False: 0] ------------------ 1156| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1157| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1158| | 1159| 67| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (1159:20): [True: 67, False: 0] ------------------ 1160| 67| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1161| 67| (uint16_t)pos1, &type1); 1162| 67| c1 = get_copy_of_container(c1, &type1, is_cow(x1)); 1163| 67| if (is_cow(x1)) { ------------------ | Branch (1163:17): [True: 0, False: 67] ------------------ 1164| 0| ra_set_container_at_index(&x1->high_low_container, pos1, c1, 1165| 0| type1); 1166| 0| } 1167| 67| ra_append(&answer->high_low_container, s1, c1, type1); 1168| 67| pos1++; 1169| 67| if (pos1 == length1) break; ------------------ | Branch (1169:17): [True: 67, False: 0] ------------------ 1170| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1171| | 1172| 0| } else { // s1 > s2 1173| 0| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 1174| 0| (uint16_t)pos2, &type2); 1175| 0| c2 = get_copy_of_container(c2, &type2, is_cow(x2)); 1176| 0| if (is_cow(x2)) { ------------------ | Branch (1176:17): [True: 0, False: 0] ------------------ 1177| 0| ra_set_container_at_index(&x2->high_low_container, pos2, c2, 1178| 0| type2); 1179| 0| } 1180| 0| ra_append(&answer->high_low_container, s2, c2, type2); 1181| 0| pos2++; 1182| 0| if (pos2 == length2) break; ------------------ | Branch (1182:17): [True: 0, False: 0] ------------------ 1183| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1184| 0| } 1185| 6.19k| } 1186| 6.19k| if (pos1 == length1) { ------------------ | Branch (1186:9): [True: 6.19k, False: 0] ------------------ 1187| 6.19k| ra_append_copy_range(&answer->high_low_container, 1188| 6.19k| &x2->high_low_container, pos2, length2, 1189| 6.19k| is_cow(x2)); 1190| 6.19k| } else if (pos2 == length2) { ------------------ | Branch (1190:16): [True: 0, False: 0] ------------------ 1191| 0| ra_append_copy_range(&answer->high_low_container, 1192| 0| &x1->high_low_container, pos1, length1, 1193| 0| is_cow(x1)); 1194| 0| } 1195| 6.19k| return answer; 1196| 6.24k|} roaring_bitmap_xor_inplace: 1201| 6.24k| const roaring_bitmap_t *x2) { 1202| 6.24k| assert(x1 != x2); 1203| 6.24k| uint8_t result_type = 0; 1204| 6.24k| int length1 = x1->high_low_container.size; 1205| 6.24k| const int length2 = x2->high_low_container.size; 1206| | 1207| 6.24k| if (0 == length2) return; ------------------ | Branch (1207:9): [True: 14, False: 6.23k] ------------------ 1208| | 1209| 6.23k| if (0 == length1) { ------------------ | Branch (1209:9): [True: 1.50k, False: 4.73k] ------------------ 1210| 1.50k| roaring_bitmap_overwrite(x1, x2); 1211| 1.50k| return; 1212| 1.50k| } 1213| | 1214| | // XOR can have new containers inserted from x2, but can also 1215| | // lose containers when x1 and x2 are nonempty and identical. 1216| | 1217| 4.73k| int pos1 = 0, pos2 = 0; 1218| 4.73k| uint8_t type1, type2; 1219| 4.73k| uint16_t s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1220| 4.73k| uint16_t s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1221| 6.71k| while (true) { ------------------ | Branch (1221:12): [True: 6.71k, Folded] ------------------ 1222| 6.71k| if (s1 == s2) { ------------------ | Branch (1222:13): [True: 6.61k, False: 102] ------------------ 1223| 6.61k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1224| 6.61k| (uint16_t)pos1, &type1); 1225| 6.61k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 1226| 6.61k| (uint16_t)pos2, &type2); 1227| | 1228| | // We do the computation "in place" only when c1 is not a shared 1229| | // container. Rationale: using a shared container safely with in 1230| | // place computation would require making a copy and then doing the 1231| | // computation in place which is likely less efficient than avoiding 1232| | // in place entirely and always generating a new container. 1233| | 1234| 6.61k| container_t *c; 1235| 6.61k| if (type1 == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 6.61k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1235:17): [True: 0, False: 6.61k] ------------------ 1236| 0| c = container_xor(c1, type1, c2, type2, &result_type); 1237| 0| shared_container_free(CAST_shared(c1)); // so release ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1238| 6.61k| } else { 1239| 6.61k| c = container_ixor(c1, type1, c2, type2, &result_type); 1240| 6.61k| } 1241| | 1242| 6.61k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (1242:17): [True: 4.43k, False: 2.17k] ------------------ 1243| 4.43k| ra_set_container_at_index(&x1->high_low_container, pos1, c, 1244| 4.43k| result_type); 1245| 4.43k| ++pos1; 1246| 4.43k| } else { 1247| 2.17k| container_free(c, result_type); 1248| 2.17k| roaring_inplace_merge_bulk(x1, x2, pos1, pos1 + 1, pos2 + 1, 1249| 2.17k| true); 1250| 2.17k| return; 1251| 2.17k| } 1252| | 1253| 4.43k| ++pos2; 1254| 4.43k| if (pos1 == length1) break; ------------------ | Branch (1254:17): [True: 2.45k, False: 1.98k] ------------------ 1255| 1.98k| if (pos2 == length2) break; ------------------ | Branch (1255:17): [True: 0, False: 1.98k] ------------------ 1256| 1.98k| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1257| 1.98k| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1258| | 1259| 1.98k| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (1259:20): [True: 0, False: 102] ------------------ 1260| 0| pos1++; 1261| 0| if (pos1 == length1) break; ------------------ | Branch (1261:17): [True: 0, False: 0] ------------------ 1262| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1263| | 1264| 102| } else { // s1 > s2 1265| 102| roaring_inplace_merge_bulk(x1, x2, pos1, pos1, pos2, true); 1266| 102| return; 1267| 102| } 1268| 6.71k| } 1269| 2.45k| if (pos1 == length1) { ------------------ | Branch (1269:9): [True: 2.45k, False: 0] ------------------ 1270| 2.45k| ra_append_copy_range(&x1->high_low_container, &x2->high_low_container, 1271| 2.45k| pos2, length2, is_cow(x2)); 1272| 2.45k| } 1273| 2.45k|} roaring_bitmap_andnot: 1276| 6.24k| const roaring_bitmap_t *x2) { 1277| 6.24k| uint8_t result_type = 0; 1278| 6.24k| const int length1 = x1->high_low_container.size, 1279| 6.24k| length2 = x2->high_low_container.size; 1280| 6.24k| if (0 == length1) { ------------------ | Branch (1280:9): [True: 0, False: 6.24k] ------------------ 1281| 0| roaring_bitmap_t *empty_bitmap = roaring_bitmap_create(); 1282| 0| roaring_bitmap_set_copy_on_write(empty_bitmap, 1283| 0| is_cow(x1) || is_cow(x2)); ------------------ | Branch (1283:42): [True: 0, False: 0] | Branch (1283:56): [True: 0, False: 0] ------------------ 1284| 0| return empty_bitmap; 1285| 0| } 1286| 6.24k| if (0 == length2) { ------------------ | Branch (1286:9): [True: 51, False: 6.19k] ------------------ 1287| 51| return roaring_bitmap_copy(x1); 1288| 51| } 1289| 6.19k| roaring_bitmap_t *answer = roaring_bitmap_create_with_capacity(length1); 1290| 6.19k| roaring_bitmap_set_copy_on_write(answer, is_cow(x1) || is_cow(x2)); ------------------ | Branch (1290:46): [True: 0, False: 6.19k] | Branch (1290:60): [True: 0, False: 6.19k] ------------------ 1291| | 1292| 6.19k| int pos1 = 0, pos2 = 0; 1293| 6.19k| uint8_t type1, type2; 1294| 6.19k| uint16_t s1 = 0; 1295| 6.19k| uint16_t s2 = 0; 1296| 6.19k| while (true) { ------------------ | Branch (1296:12): [True: 6.19k, Folded] ------------------ 1297| 6.19k| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1298| 6.19k| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1299| | 1300| 6.19k| if (s1 == s2) { ------------------ | Branch (1300:13): [True: 6.12k, False: 67] ------------------ 1301| 6.12k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1302| 6.12k| (uint16_t)pos1, &type1); 1303| 6.12k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 1304| 6.12k| (uint16_t)pos2, &type2); 1305| 6.12k| container_t *c = 1306| 6.12k| container_andnot(c1, type1, c2, type2, &result_type); 1307| | 1308| 6.12k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (1308:17): [True: 5.92k, False: 205] ------------------ 1309| 5.92k| ra_append(&answer->high_low_container, s1, c, result_type); 1310| 5.92k| } else { 1311| 205| container_free(c, result_type); 1312| 205| } 1313| 6.12k| ++pos1; 1314| 6.12k| ++pos2; 1315| 6.12k| if (pos1 == length1) break; ------------------ | Branch (1315:17): [True: 6.12k, False: 0] ------------------ 1316| 0| if (pos2 == length2) break; ------------------ | Branch (1316:17): [True: 0, False: 0] ------------------ 1317| 67| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (1317:20): [True: 67, False: 0] ------------------ 1318| 67| const int next_pos1 = 1319| 67| ra_advance_until(&x1->high_low_container, s2, pos1); 1320| 67| ra_append_copy_range(&answer->high_low_container, 1321| 67| &x1->high_low_container, pos1, next_pos1, 1322| 67| is_cow(x1)); 1323| | // TODO : perhaps some of the copy_on_write should be based on 1324| | // answer rather than x1 (more stringent?). Many similar cases 1325| 67| pos1 = next_pos1; 1326| 67| if (pos1 == length1) break; ------------------ | Branch (1326:17): [True: 67, False: 0] ------------------ 1327| 67| } else { // s1 > s2 1328| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 1329| 0| if (pos2 == length2) break; ------------------ | Branch (1329:17): [True: 0, False: 0] ------------------ 1330| 0| } 1331| 6.19k| } 1332| 6.19k| if (pos2 == length2) { ------------------ | Branch (1332:9): [True: 3.99k, False: 2.20k] ------------------ 1333| 3.99k| ra_append_copy_range(&answer->high_low_container, 1334| 3.99k| &x1->high_low_container, pos1, length1, 1335| 3.99k| is_cow(x1)); 1336| 3.99k| } 1337| 6.19k| return answer; 1338| 6.24k|} roaring_bitmap_andnot_inplace: 1343| 6.24k| const roaring_bitmap_t *x2) { 1344| 6.24k| assert(x1 != x2); 1345| | 1346| 6.24k| uint8_t result_type = 0; 1347| 6.24k| int length1 = x1->high_low_container.size; 1348| 6.24k| const int length2 = x2->high_low_container.size; 1349| 6.24k| int intersection_size = 0; 1350| | 1351| 6.24k| if (0 == length2) return; ------------------ | Branch (1351:9): [True: 434, False: 5.81k] ------------------ 1352| | 1353| 5.81k| if (0 == length1) { ------------------ | Branch (1353:9): [True: 0, False: 5.81k] ------------------ 1354| 0| roaring_bitmap_clear(x1); 1355| 0| return; 1356| 0| } 1357| | 1358| 5.81k| int pos1 = 0, pos2 = 0; 1359| 5.81k| uint8_t type1, type2; 1360| 5.81k| uint16_t s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1361| 5.81k| uint16_t s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1362| 5.81k| while (true) { ------------------ | Branch (1362:12): [True: 5.81k, Folded] ------------------ 1363| 5.81k| if (s1 == s2) { ------------------ | Branch (1363:13): [True: 5.81k, False: 0] ------------------ 1364| 5.81k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 1365| 5.81k| (uint16_t)pos1, &type1); 1366| 5.81k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 1367| 5.81k| (uint16_t)pos2, &type2); 1368| | 1369| | // We do the computation "in place" only when c1 is not a shared 1370| | // container. Rationale: using a shared container safely with in 1371| | // place computation would require making a copy and then doing the 1372| | // computation in place which is likely less efficient than avoiding 1373| | // in place entirely and always generating a new container. 1374| | 1375| 5.81k| container_t *c; 1376| 5.81k| if (type1 == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 5.81k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1376:17): [True: 0, False: 5.81k] ------------------ 1377| 0| c = container_andnot(c1, type1, c2, type2, &result_type); 1378| 0| shared_container_free(CAST_shared(c1)); // release ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1379| 5.81k| } else { 1380| 5.81k| c = container_iandnot(c1, type1, c2, type2, &result_type); 1381| 5.81k| } 1382| | 1383| 5.81k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (1383:17): [True: 4.31k, False: 1.50k] ------------------ 1384| 4.31k| ra_replace_key_and_container_at_index(&x1->high_low_container, 1385| 4.31k| intersection_size++, s1, 1386| 4.31k| c, result_type); 1387| 4.31k| } else { 1388| 1.50k| container_free(c, result_type); 1389| 1.50k| } 1390| | 1391| 5.81k| ++pos1; 1392| 5.81k| ++pos2; 1393| 5.81k| if (pos1 == length1) break; ------------------ | Branch (1393:17): [True: 3.77k, False: 2.03k] ------------------ 1394| 2.03k| if (pos2 == length2) break; ------------------ | Branch (1394:17): [True: 2.03k, False: 0] ------------------ 1395| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1396| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1397| | 1398| 0| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (1398:20): [True: 0, False: 0] ------------------ 1399| 0| if (pos1 != intersection_size) { ------------------ | Branch (1399:17): [True: 0, False: 0] ------------------ 1400| 0| container_t *c1 = ra_get_container_at_index( 1401| 0| &x1->high_low_container, (uint16_t)pos1, &type1); 1402| | 1403| 0| ra_replace_key_and_container_at_index( 1404| 0| &x1->high_low_container, intersection_size, s1, c1, type1); 1405| 0| } 1406| 0| intersection_size++; 1407| 0| pos1++; 1408| 0| if (pos1 == length1) break; ------------------ | Branch (1408:17): [True: 0, False: 0] ------------------ 1409| 0| s1 = ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 1410| | 1411| 0| } else { // s1 > s2 1412| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 1413| 0| if (pos2 == length2) break; ------------------ | Branch (1413:17): [True: 0, False: 0] ------------------ 1414| 0| s2 = ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 1415| 0| } 1416| 5.81k| } 1417| | 1418| 5.81k| if (pos1 < length1) { ------------------ | Branch (1418:9): [True: 2.03k, False: 3.77k] ------------------ 1419| | // all containers between intersection_size and 1420| | // pos1 are junk. However, they have either been moved 1421| | // (thus still referenced) or involved in an iandnot 1422| | // that will clean up all containers that could not be reused. 1423| | // Thus we should not free the junk containers between 1424| | // intersection_size and pos1. 1425| 2.03k| if (pos1 > intersection_size) { ------------------ | Branch (1425:13): [True: 38, False: 1.99k] ------------------ 1426| | // left slide of remaining items 1427| 38| ra_copy_range(&x1->high_low_container, pos1, length1, 1428| 38| intersection_size); 1429| 38| } 1430| | // else current placement is fine 1431| 2.03k| intersection_size += (length1 - pos1); 1432| 2.03k| } 1433| 5.81k| ra_downsize(&x1->high_low_container, intersection_size); 1434| 5.81k|} roaring_bitmap_get_cardinality: 1436| 62.4k|uint64_t roaring_bitmap_get_cardinality(const roaring_bitmap_t *r) { 1437| 62.4k| const roaring_array_t *ra = &r->high_low_container; 1438| | 1439| 62.4k| uint64_t card = 0; 1440| 409k| for (int i = 0; i < ra->size; ++i) ------------------ | Branch (1440:21): [True: 347k, False: 62.4k] ------------------ 1441| 347k| card += container_get_cardinality(ra->containers[i], ra->typecodes[i]); 1442| 62.4k| return card; 1443| 62.4k|} roaring_bitmap_is_empty: 1506| 12.4k|bool roaring_bitmap_is_empty(const roaring_bitmap_t *r) { 1507| 12.4k| return r->high_low_container.size == 0; 1508| 12.4k|} roaring_bitmap_to_uint32_array: 1510| 6.24k|void roaring_bitmap_to_uint32_array(const roaring_bitmap_t *r, uint32_t *ans) { 1511| 6.24k| ra_to_uint32_array(&r->high_low_container, ans); 1512| 6.24k|} roaring_bitmap_run_optimize: 1530| 12.4k|bool roaring_bitmap_run_optimize(roaring_bitmap_t *r) { 1531| 12.4k| bool answer = false; 1532| 24.9k| for (int i = 0; i < r->high_low_container.size; i++) { ------------------ | Branch (1532:21): [True: 12.4k, False: 12.4k] ------------------ 1533| 12.4k| uint8_t type_original, type_after; 1534| 12.4k| ra_unshare_container_at_index( 1535| 12.4k| &r->high_low_container, 1536| 12.4k| (uint16_t)i); // TODO: this introduces extra cloning! 1537| 12.4k| container_t *c = ra_get_container_at_index(&r->high_low_container, 1538| 12.4k| (uint16_t)i, &type_original); 1539| 12.4k| container_t *c1 = convert_run_optimize(c, type_original, &type_after); 1540| 12.4k| if (type_after == RUN_CONTAINER_TYPE) { ------------------ | | 50| 12.4k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (1540:13): [True: 3.97k, False: 8.51k] ------------------ 1541| | answer = true; 1542| 3.97k| } 1543| 12.4k| ra_set_container_at_index(&r->high_low_container, i, c1, type_after); 1544| 12.4k| } 1545| 12.4k| return answer; 1546| 12.4k|} roaring_bitmap_shrink_to_fit: 1548| 6.24k|size_t roaring_bitmap_shrink_to_fit(roaring_bitmap_t *r) { 1549| 6.24k| size_t answer = 0; 1550| 12.4k| for (int i = 0; i < r->high_low_container.size; i++) { ------------------ | Branch (1550:21): [True: 6.24k, False: 6.24k] ------------------ 1551| 6.24k| uint8_t type_original; 1552| 6.24k| container_t *c = ra_get_container_at_index(&r->high_low_container, 1553| 6.24k| (uint16_t)i, &type_original); 1554| 6.24k| answer += container_shrink_to_fit(c, type_original); 1555| 6.24k| } 1556| 6.24k| answer += ra_shrink_to_fit(&r->high_low_container); 1557| 6.24k| return answer; 1558| 6.24k|} roaring_bitmap_remove_run_compression: 1564| 6.24k|bool roaring_bitmap_remove_run_compression(roaring_bitmap_t *r) { 1565| 6.24k| bool answer = false; 1566| 80.0k| for (int i = 0; i < r->high_low_container.size; i++) { ------------------ | Branch (1566:21): [True: 73.8k, False: 6.24k] ------------------ 1567| 73.8k| uint8_t type_original, type_after; 1568| 73.8k| container_t *c = ra_get_container_at_index(&r->high_low_container, 1569| 73.8k| (uint16_t)i, &type_original); 1570| 73.8k| if (get_container_type(c, type_original) == RUN_CONTAINER_TYPE) { ------------------ | | 50| 73.8k|#define RUN_CONTAINER_TYPE 3 ------------------ | Branch (1570:13): [True: 66.5k, False: 7.24k] ------------------ 1571| 66.5k| answer = true; 1572| 66.5k| if (type_original == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 66.5k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1572:17): [True: 0, False: 66.5k] ------------------ 1573| 0| run_container_t *truec = CAST_run(CAST_shared(c)->container); ------------------ | | 77| 0|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1574| 0| int32_t card = run_container_cardinality(truec); 1575| 0| container_t *c1 = convert_to_bitset_or_array_container( 1576| 0| truec, card, &type_after); 1577| 0| shared_container_free(CAST_shared(c)); // frees run as needed ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1578| 0| ra_set_container_at_index(&r->high_low_container, i, c1, 1579| 0| type_after); 1580| | 1581| 66.5k| } else { 1582| 66.5k| int32_t card = run_container_cardinality(CAST_run(c)); ------------------ | | 77| 66.5k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 66.5k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1583| 66.5k| container_t *c1 = convert_to_bitset_or_array_container( 1584| 66.5k| CAST_run(c), card, &type_after); ------------------ | | 77| 66.5k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 66.5k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1585| 66.5k| run_container_free(CAST_run(c)); ------------------ | | 77| 66.5k|#define CAST_run(c) CAST(run_container_t *, c) // safer downcast | | ------------------ | | | | 86| 66.5k|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1586| 66.5k| ra_set_container_at_index(&r->high_low_container, i, c1, 1587| 66.5k| type_after); 1588| 66.5k| } 1589| 66.5k| } 1590| 73.8k| } 1591| 6.24k| return answer; 1592| 6.24k|} roaring_bitmap_portable_size_in_bytes: 1627| 12.4k|size_t roaring_bitmap_portable_size_in_bytes(const roaring_bitmap_t *r) { 1628| 12.4k| return ra_portable_size_in_bytes(&r->high_low_container); 1629| 12.4k|} roaring_bitmap_portable_deserialize_safe: 1632| 12.4k| size_t maxbytes) { 1633| 12.4k| roaring_bitmap_t *ans = 1634| 12.4k| (roaring_bitmap_t *)roaring_malloc(sizeof(roaring_bitmap_t)); 1635| 12.4k| if (ans == NULL) { ------------------ | Branch (1635:9): [True: 0, False: 12.4k] ------------------ 1636| 0| return NULL; 1637| 0| } 1638| 12.4k| size_t bytesread; 1639| 12.4k| bool is_ok = ra_portable_deserialize(&ans->high_low_container, buf, 1640| 12.4k| maxbytes, &bytesread); 1641| 12.4k| if (!is_ok) { ------------------ | Branch (1641:9): [True: 6.23k, False: 6.25k] ------------------ 1642| 6.23k| roaring_free(ans); 1643| 6.23k| return NULL; 1644| 6.23k| } 1645| 6.25k| roaring_bitmap_set_copy_on_write(ans, false); 1646| 6.25k| if (!is_ok) { ------------------ | Branch (1646:9): [True: 0, False: 6.25k] ------------------ 1647| 0| roaring_free(ans); 1648| 0| return NULL; 1649| 0| } 1650| 6.25k| return ans; 1651| 6.25k|} roaring_bitmap_portable_serialize: 1662| 6.24k|size_t roaring_bitmap_portable_serialize(const roaring_bitmap_t *r, char *buf) { 1663| 6.24k| return ra_portable_serialize(&r->high_low_container, buf); 1664| 6.24k|} roaring_iterate: 1745| 6.23k| void *ptr) { 1746| 6.23k| const roaring_array_t *ra = &r->high_low_container; 1747| | 1748| 69.6k| for (int i = 0; i < ra->size; ++i) ------------------ | Branch (1748:21): [True: 63.3k, False: 6.23k] ------------------ 1749| 63.3k| if (!container_iterate(ra->containers[i], ra->typecodes[i], ------------------ | Branch (1749:13): [True: 0, False: 63.3k] ------------------ 1750| 63.3k| ((uint32_t)ra->keys[i]) << 16, iterator, ptr)) { 1751| 0| return false; 1752| 0| } 1753| 6.23k| return true; 1754| 6.23k|} roaring_iterator_init: 1861| 12.4k| roaring_uint32_iterator_t *newit) { 1862| 12.4k| newit->parent = r; 1863| 12.4k| newit->container_index = 0; 1864| 12.4k| newit->has_value = loadfirstvalue(newit); 1865| 12.4k|} roaring_uint32_iterator_move_equalorlarger: 1893| 6.24k| uint32_t val) { 1894| 6.24k| uint16_t hb = val >> 16; 1895| 6.24k| const int i = ra_get_index(&it->parent->high_low_container, hb); 1896| 6.24k| if (i >= 0) { ------------------ | Branch (1896:9): [True: 6.18k, False: 57] ------------------ 1897| 6.18k| uint32_t lowvalue = 1898| 6.18k| container_maximum(it->parent->high_low_container.containers[i], 1899| 6.18k| it->parent->high_low_container.typecodes[i]); 1900| 6.18k| uint16_t lb = val & 0xFFFF; 1901| 6.18k| if (lowvalue < lb) { ------------------ | Branch (1901:13): [True: 19, False: 6.17k] ------------------ 1902| | // will have to load first value of next container 1903| 19| it->container_index = i + 1; 1904| 6.17k| } else { 1905| | // the value is necessarily within the range of the container 1906| 6.17k| it->container_index = i; 1907| 6.17k| it->has_value = loadfirstvalue_largeorequal(it, val); 1908| 6.17k| return it->has_value; 1909| 6.17k| } 1910| 6.18k| } else { 1911| | // there is no matching, so we are going for the next container 1912| 57| it->container_index = -i - 1; 1913| 57| } 1914| 76| it->has_value = loadfirstvalue(it); 1915| 76| return it->has_value; 1916| 6.24k|} roaring_uint32_iterator_advance: 1918| 329k|bool roaring_uint32_iterator_advance(roaring_uint32_iterator_t *it) { 1919| 329k| if (it->container_index >= it->parent->high_low_container.size) { ------------------ | Branch (1919:9): [True: 0, False: 329k] ------------------ 1920| 0| return (it->has_value = false); 1921| 0| } 1922| 329k| if (it->container_index < 0) { ------------------ | Branch (1922:9): [True: 0, False: 329k] ------------------ 1923| 0| it->container_index = 0; 1924| 0| return (it->has_value = loadfirstvalue(it)); 1925| 0| } 1926| 329k| uint16_t low16 = (uint16_t)it->current_value; 1927| 329k| if (container_iterator_next(it->container, it->typecode, &it->container_it, ------------------ | Branch (1927:9): [True: 323k, False: 6.04k] ------------------ 1928| 329k| &low16)) { 1929| 323k| it->current_value = it->highbits | low16; 1930| 323k| return (it->has_value = true); 1931| 323k| } 1932| 6.04k| it->container_index++; 1933| 6.04k| return (it->has_value = loadfirstvalue(it)); 1934| 329k|} roaring_bitmap_equals: 2129| 12.4k| const roaring_bitmap_t *r2) { 2130| 12.4k| const roaring_array_t *ra1 = &r1->high_low_container; 2131| 12.4k| const roaring_array_t *ra2 = &r2->high_low_container; 2132| | 2133| 12.4k| if (ra1->size != ra2->size) { ------------------ | Branch (2133:9): [True: 2.31k, False: 10.1k] ------------------ 2134| 2.31k| return false; 2135| 2.31k| } 2136| 87.8k| for (int i = 0; i < ra1->size; ++i) { ------------------ | Branch (2136:21): [True: 77.7k, False: 10.1k] ------------------ 2137| 77.7k| if (ra1->keys[i] != ra2->keys[i]) { ------------------ | Branch (2137:13): [True: 25, False: 77.7k] ------------------ 2138| 25| return false; 2139| 25| } 2140| 77.7k| } 2141| 83.9k| for (int i = 0; i < ra1->size; ++i) { ------------------ | Branch (2141:21): [True: 77.7k, False: 6.24k] ------------------ 2142| 77.7k| bool areequal = container_equals(ra1->containers[i], ra1->typecodes[i], 2143| 77.7k| ra2->containers[i], ra2->typecodes[i]); 2144| 77.7k| if (!areequal) { ------------------ | Branch (2144:13): [True: 3.90k, False: 73.8k] ------------------ 2145| 3.90k| return false; 2146| 3.90k| } 2147| 77.7k| } 2148| 6.24k| return true; 2149| 10.1k|} roaring_bitmap_is_subset: 2152| 11.2k| const roaring_bitmap_t *r2) { 2153| 11.2k| const roaring_array_t *ra1 = &r1->high_low_container; 2154| 11.2k| const roaring_array_t *ra2 = &r2->high_low_container; 2155| | 2156| 11.2k| const int length1 = ra1->size, length2 = ra2->size; 2157| | 2158| 11.2k| int pos1 = 0, pos2 = 0; 2159| | 2160| 11.2k| while (pos1 < length1 && pos2 < length2) { ------------------ | Branch (2160:12): [True: 10.8k, False: 410] | Branch (2160:30): [True: 10.8k, False: 51] ------------------ 2161| 10.8k| const uint16_t s1 = ra_get_key_at_index(ra1, (uint16_t)pos1); 2162| 10.8k| const uint16_t s2 = ra_get_key_at_index(ra2, (uint16_t)pos2); 2163| | 2164| 10.8k| if (s1 == s2) { ------------------ | Branch (2164:13): [True: 10.7k, False: 103] ------------------ 2165| 10.7k| uint8_t type1, type2; 2166| 10.7k| container_t *c1 = 2167| 10.7k| ra_get_container_at_index(ra1, (uint16_t)pos1, &type1); 2168| 10.7k| container_t *c2 = 2169| 10.7k| ra_get_container_at_index(ra2, (uint16_t)pos2, &type2); 2170| 10.7k| if (!container_is_subset(c1, type1, c2, type2)) return false; ------------------ | Branch (2170:17): [True: 10.7k, False: 0] ------------------ 2171| 0| ++pos1; 2172| 0| ++pos2; 2173| 103| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (2173:20): [True: 103, False: 0] ------------------ 2174| 103| return false; 2175| 103| } else { // s1 > s2 2176| 0| pos2 = ra_advance_until(ra2, s1, pos2); 2177| 0| } 2178| 10.8k| } 2179| 461| if (pos1 == length1) ------------------ | Branch (2179:9): [True: 410, False: 51] ------------------ 2180| 410| return true; 2181| 51| else 2182| 51| return false; 2183| 461|} roaring_bitmap_flip_inplace: 2352| 6.24k| uint64_t range_end) { 2353| 6.24k| if (range_start >= range_end || range_start > (uint64_t)UINT32_MAX + 1) { ------------------ | Branch (2353:9): [True: 5.72k, False: 522] | Branch (2353:37): [True: 0, False: 522] ------------------ 2354| 5.72k| return; 2355| 5.72k| } 2356| 522| roaring_bitmap_flip_inplace_closed(x1, (uint32_t)range_start, 2357| 522| (uint32_t)(range_end - 1)); 2358| 522|} roaring_bitmap_flip_inplace_closed: 2362| 6.76k| uint32_t range_end) { 2363| 6.76k| if (range_start > range_end) { ------------------ | Branch (2363:9): [True: 156, False: 6.61k] ------------------ 2364| 156| return; // empty range 2365| 156| } 2366| | 2367| 6.61k| uint16_t hb_start = (uint16_t)(range_start >> 16); 2368| 6.61k| const uint16_t lb_start = (uint16_t)range_start; 2369| 6.61k| uint16_t hb_end = (uint16_t)(range_end >> 16); 2370| 6.61k| const uint16_t lb_end = (uint16_t)range_end; 2371| | 2372| 6.61k| if (hb_start == hb_end) { ------------------ | Branch (2372:9): [True: 6.16k, False: 446] ------------------ 2373| 6.16k| inplace_flip_container(&x1->high_low_container, hb_start, lb_start, 2374| 6.16k| lb_end); 2375| 6.16k| } else { 2376| | // start and end containers are distinct 2377| 446| if (lb_start > 0) { ------------------ | Branch (2377:13): [True: 359, False: 87] ------------------ 2378| | // handle first (partial) container 2379| 359| inplace_flip_container(&x1->high_low_container, hb_start, lb_start, 2380| 359| 0xFFFF); 2381| 359| ++hb_start; // for the full containers. Can't wrap. 2382| 359| } 2383| | 2384| 446| if (lb_end != 0xFFFF) --hb_end; ------------------ | Branch (2384:13): [True: 429, False: 17] ------------------ 2385| | 2386| 23.6k| for (uint32_t hb = hb_start; hb <= hb_end; ++hb) { ------------------ | Branch (2386:38): [True: 23.2k, False: 446] ------------------ 2387| 23.2k| inplace_fully_flip_container(&x1->high_low_container, (uint16_t)hb); 2388| 23.2k| } 2389| | // handle a partial final container 2390| 446| if (lb_end != 0xFFFF) { ------------------ | Branch (2390:13): [True: 429, False: 17] ------------------ 2391| 429| inplace_flip_container(&x1->high_low_container, hb_end + 1, 0, 2392| 429| lb_end); 2393| 429| ++hb_end; 2394| 429| } 2395| 446| } 2396| 6.61k|} roaring_bitmap_rank: 2862| 6.24k|uint64_t roaring_bitmap_rank(const roaring_bitmap_t *bm, uint32_t x) { 2863| 6.24k| uint64_t size = 0; 2864| 6.24k| uint32_t xhigh = x >> 16; 2865| 6.24k| for (int i = 0; i < bm->high_low_container.size; i++) { ------------------ | Branch (2865:21): [True: 6.24k, False: 0] ------------------ 2866| 6.24k| uint32_t key = bm->high_low_container.keys[i]; 2867| 6.24k| if (xhigh > key) { ------------------ | Branch (2867:13): [True: 0, False: 6.24k] ------------------ 2868| 0| size += 2869| 0| container_get_cardinality(bm->high_low_container.containers[i], 2870| 0| bm->high_low_container.typecodes[i]); 2871| 6.24k| } else if (xhigh == key) { ------------------ | Branch (2871:20): [True: 6.24k, False: 0] ------------------ 2872| 6.24k| return size + container_rank(bm->high_low_container.containers[i], 2873| 6.24k| bm->high_low_container.typecodes[i], 2874| 6.24k| x & 0xFFFF); 2875| 6.24k| } else { 2876| 0| return size; 2877| 0| } 2878| 6.24k| } 2879| 0| return size; 2880| 6.24k|} roaring_bitmap_minimum: 2945| 6.24k|uint32_t roaring_bitmap_minimum(const roaring_bitmap_t *bm) { 2946| 6.24k| if (bm->high_low_container.size > 0) { ------------------ | Branch (2946:9): [True: 6.19k, False: 51] ------------------ 2947| 6.19k| container_t *c = bm->high_low_container.containers[0]; 2948| 6.19k| uint8_t type = bm->high_low_container.typecodes[0]; 2949| 6.19k| uint32_t key = bm->high_low_container.keys[0]; 2950| 6.19k| uint32_t lowvalue = container_minimum(c, type); 2951| 6.19k| return lowvalue | (key << 16); 2952| 6.19k| } 2953| 51| return UINT32_MAX; 2954| 6.24k|} roaring_bitmap_maximum: 2960| 6.24k|uint32_t roaring_bitmap_maximum(const roaring_bitmap_t *bm) { 2961| 6.24k| if (bm->high_low_container.size > 0) { ------------------ | Branch (2961:9): [True: 6.19k, False: 51] ------------------ 2962| 6.19k| container_t *container = 2963| 6.19k| bm->high_low_container.containers[bm->high_low_container.size - 1]; 2964| 6.19k| uint8_t typecode = 2965| 6.19k| bm->high_low_container.typecodes[bm->high_low_container.size - 1]; 2966| 6.19k| uint32_t key = 2967| 6.19k| bm->high_low_container.keys[bm->high_low_container.size - 1]; 2968| 6.19k| uint32_t lowvalue = container_maximum(container, typecode); 2969| 6.19k| return lowvalue | (key << 16); 2970| 6.19k| } 2971| 51| return 0; 2972| 6.24k|} roaring_bitmap_select: 2975| 6.24k| uint32_t *element) { 2976| 6.24k| container_t *container; 2977| 6.24k| uint8_t typecode; 2978| 6.24k| uint16_t key; 2979| 6.24k| uint32_t start_rank = 0; 2980| 6.24k| int i = 0; 2981| 6.24k| bool valid = false; 2982| 12.4k| while (!valid && i < bm->high_low_container.size) { ------------------ | Branch (2982:12): [True: 6.68k, False: 5.80k] | Branch (2982:22): [True: 6.24k, False: 438] ------------------ 2983| 6.24k| container = bm->high_low_container.containers[i]; 2984| 6.24k| typecode = bm->high_low_container.typecodes[i]; 2985| 6.24k| valid = 2986| 6.24k| container_select(container, typecode, &start_rank, rank, element); 2987| 6.24k| i++; 2988| 6.24k| } 2989| | 2990| 6.24k| if (valid) { ------------------ | Branch (2990:9): [True: 5.80k, False: 438] ------------------ 2991| 5.80k| key = bm->high_low_container.keys[i - 1]; 2992| 5.80k| *element |= (((uint32_t)key) << 16); // w/o cast, key promotes signed 2993| 5.80k| return true; 2994| 5.80k| } else 2995| 438| return false; 2996| 6.24k|} roaring_bitmap_intersect: 2999| 6.24k| const roaring_bitmap_t *x2) { 3000| 6.24k| const int length1 = x1->high_low_container.size, 3001| 6.24k| length2 = x2->high_low_container.size; 3002| 6.24k| uint64_t answer = 0; 3003| 6.24k| int pos1 = 0, pos2 = 0; 3004| | 3005| 6.62k| while (pos1 < length1 && pos2 < length2) { ------------------ | Branch (3005:12): [True: 6.24k, False: 383] | Branch (3005:30): [True: 6.19k, False: 51] ------------------ 3006| 6.19k| const uint16_t s1 = 3007| 6.19k| ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 3008| 6.19k| const uint16_t s2 = 3009| 6.19k| ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 3010| | 3011| 6.19k| if (s1 == s2) { ------------------ | Branch (3011:13): [True: 6.12k, False: 67] ------------------ 3012| 6.12k| uint8_t type1, type2; 3013| 6.12k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 3014| 6.12k| (uint16_t)pos1, &type1); 3015| 6.12k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 3016| 6.12k| (uint16_t)pos2, &type2); 3017| 6.12k| if (container_intersect(c1, type1, c2, type2)) return true; ------------------ | Branch (3017:17): [True: 5.81k, False: 316] ------------------ 3018| 316| ++pos1; 3019| 316| ++pos2; 3020| 316| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (3020:20): [True: 67, False: 0] ------------------ 3021| 67| pos1 = ra_advance_until(&x1->high_low_container, s2, pos1); 3022| 67| } else { // s1 > s2 3023| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 3024| 0| } 3025| 6.19k| } 3026| 434| return answer != 0; 3027| 6.24k|} roaring_bitmap_and_cardinality: 3049| 24.9k| const roaring_bitmap_t *x2) { 3050| 24.9k| const int length1 = x1->high_low_container.size, 3051| 24.9k| length2 = x2->high_low_container.size; 3052| 24.9k| uint64_t answer = 0; 3053| 24.9k| int pos1 = 0, pos2 = 0; 3054| 49.7k| while (pos1 < length1 && pos2 < length2) { ------------------ | Branch (3054:12): [True: 24.9k, False: 24.7k] | Branch (3054:30): [True: 24.7k, False: 204] ------------------ 3055| 24.7k| const uint16_t s1 = 3056| 24.7k| ra_get_key_at_index(&x1->high_low_container, (uint16_t)pos1); 3057| 24.7k| const uint16_t s2 = 3058| 24.7k| ra_get_key_at_index(&x2->high_low_container, (uint16_t)pos2); 3059| | 3060| 24.7k| if (s1 == s2) { ------------------ | Branch (3060:13): [True: 24.5k, False: 268] ------------------ 3061| 24.5k| uint8_t type1, type2; 3062| 24.5k| container_t *c1 = ra_get_container_at_index(&x1->high_low_container, 3063| 24.5k| (uint16_t)pos1, &type1); 3064| 24.5k| container_t *c2 = ra_get_container_at_index(&x2->high_low_container, 3065| 24.5k| (uint16_t)pos2, &type2); 3066| 24.5k| answer += container_and_cardinality(c1, type1, c2, type2); 3067| 24.5k| ++pos1; 3068| 24.5k| ++pos2; 3069| 24.5k| } else if (s1 < s2) { // s1 < s2 ------------------ | Branch (3069:20): [True: 268, False: 0] ------------------ 3070| 268| pos1 = ra_advance_until(&x1->high_low_container, s2, pos1); 3071| 268| } else { // s1 > s2 3072| 0| pos2 = ra_advance_until(&x2->high_low_container, s1, pos2); 3073| 0| } 3074| 24.7k| } 3075| 24.9k| return answer; 3076| 24.9k|} roaring_bitmap_jaccard_index: 3079| 6.24k| const roaring_bitmap_t *x2) { 3080| 6.24k| const uint64_t c1 = roaring_bitmap_get_cardinality(x1); 3081| 6.24k| const uint64_t c2 = roaring_bitmap_get_cardinality(x2); 3082| 6.24k| const uint64_t inter = roaring_bitmap_and_cardinality(x1, x2); 3083| 6.24k| return (double)inter / (double)(c1 + c2 - inter); 3084| 6.24k|} roaring_bitmap_or_cardinality: 3087| 6.24k| const roaring_bitmap_t *x2) { 3088| 6.24k| const uint64_t c1 = roaring_bitmap_get_cardinality(x1); 3089| 6.24k| const uint64_t c2 = roaring_bitmap_get_cardinality(x2); 3090| 6.24k| const uint64_t inter = roaring_bitmap_and_cardinality(x1, x2); 3091| 6.24k| return c1 + c2 - inter; 3092| 6.24k|} roaring_bitmap_andnot_cardinality: 3095| 6.24k| const roaring_bitmap_t *x2) { 3096| 6.24k| const uint64_t c1 = roaring_bitmap_get_cardinality(x1); 3097| 6.24k| const uint64_t inter = roaring_bitmap_and_cardinality(x1, x2); 3098| 6.24k| return c1 - inter; 3099| 6.24k|} roaring_bitmap_xor_cardinality: 3102| 6.24k| const roaring_bitmap_t *x2) { 3103| 6.24k| const uint64_t c1 = roaring_bitmap_get_cardinality(x1); 3104| 6.24k| const uint64_t c2 = roaring_bitmap_get_cardinality(x2); 3105| 6.24k| const uint64_t inter = roaring_bitmap_and_cardinality(x1, x2); 3106| 6.24k| return c1 + c2 - 2 * inter; 3107| 6.24k|} roaring_bitmap_contains_range: 3114| 6.24k| uint64_t range_start, uint64_t range_end) { 3115| 6.24k| if (range_start >= range_end || range_start > (uint64_t)UINT32_MAX + 1) { ------------------ | Branch (3115:9): [True: 5.49k, False: 752] | Branch (3115:37): [True: 0, False: 752] ------------------ 3116| 5.49k| return true; 3117| 5.49k| } 3118| 752| return roaring_bitmap_contains_range_closed(r, (uint32_t)range_start, 3119| 752| (uint32_t)(range_end - 1)); 3120| 6.24k|} roaring_bitmap_contains_range_closed: 3128| 752| uint32_t range_end) { 3129| 752| if (range_start > range_end) { ------------------ | Branch (3129:9): [True: 0, False: 752] ------------------ 3130| 0| return true; 3131| 0| } // empty range are always contained! 3132| 752| if (range_end == range_start) { ------------------ | Branch (3132:9): [True: 46, False: 706] ------------------ 3133| 46| return roaring_bitmap_contains(r, (uint32_t)range_start); 3134| 46| } 3135| 706| uint16_t hb_rs = (uint16_t)(range_start >> 16); 3136| 706| uint16_t hb_re = (uint16_t)(range_end >> 16); 3137| 706| const int32_t span = hb_re - hb_rs; 3138| 706| const int32_t hlc_sz = ra_get_size(&r->high_low_container); 3139| 706| if (hlc_sz < span + 1) { ------------------ | Branch (3139:9): [True: 224, False: 482] ------------------ 3140| 224| return false; 3141| 224| } 3142| 482| int32_t is = ra_get_index(&r->high_low_container, hb_rs); 3143| 482| int32_t ie = ra_get_index(&r->high_low_container, hb_re); 3144| 482| if ((ie < 0) || (is < 0) || ((ie - is) != span) || ie >= hlc_sz) { ------------------ | Branch (3144:9): [True: 29, False: 453] | Branch (3144:21): [True: 14, False: 439] | Branch (3144:33): [True: 4, False: 435] | Branch (3144:56): [True: 0, False: 435] ------------------ 3145| 47| return false; 3146| 47| } 3147| 435| const uint32_t lb_rs = range_start & 0xFFFF; 3148| 435| const uint32_t lb_re = (range_end & 0xFFFF) + 1; 3149| 435| uint8_t type; 3150| 435| container_t *c = 3151| 435| ra_get_container_at_index(&r->high_low_container, (uint16_t)is, &type); 3152| 435| if (hb_rs == hb_re) { ------------------ | Branch (3152:9): [True: 362, False: 73] ------------------ 3153| 362| return container_contains_range(c, lb_rs, lb_re, type); 3154| 362| } 3155| 73| if (!container_contains_range(c, lb_rs, 1 << 16, type)) { ------------------ | Branch (3155:9): [True: 15, False: 58] ------------------ 3156| 15| return false; 3157| 15| } 3158| 58| c = ra_get_container_at_index(&r->high_low_container, (uint16_t)ie, &type); 3159| 58| if (!container_contains_range(c, 0, lb_re, type)) { ------------------ | Branch (3159:9): [True: 5, False: 53] ------------------ 3160| 5| return false; 3161| 5| } 3162| 787| for (int32_t i = is + 1; i < ie; ++i) { ------------------ | Branch (3162:30): [True: 756, False: 31] ------------------ 3163| 756| c = ra_get_container_at_index(&r->high_low_container, (uint16_t)i, 3164| 756| &type); 3165| 756| if (!container_is_full(c, type)) { ------------------ | Branch (3165:13): [True: 22, False: 734] ------------------ 3166| 22| return false; 3167| 22| } 3168| 756| } 3169| 31| return true; 3170| 53|} roaring_bitmap_is_strict_subset: 3173| 6.24k| const roaring_bitmap_t *r2) { 3174| 6.24k| return (roaring_bitmap_get_cardinality(r2) > ------------------ | Branch (3174:13): [True: 5.04k, False: 1.20k] ------------------ 3175| 6.24k| roaring_bitmap_get_cardinality(r1) && 3176| 5.04k| roaring_bitmap_is_subset(r1, r2)); ------------------ | Branch (3176:13): [True: 205, False: 4.83k] ------------------ 3177| 6.24k|} roaring.c:containerptr_roaring_bitmap_add: 56| 18.7k| int *index) { 57| 18.7k| roaring_array_t *ra = &r->high_low_container; 58| | 59| 18.7k| uint16_t hb = val >> 16; 60| 18.7k| const int i = ra_get_index(ra, hb); 61| 18.7k| if (i >= 0) { ------------------ | Branch (61:9): [True: 6.24k, False: 12.4k] ------------------ 62| 6.24k| ra_unshare_container_at_index(ra, (uint16_t)i); 63| 6.24k| container_t *c = ra_get_container_at_index(ra, (uint16_t)i, type); 64| 6.24k| uint8_t new_type = *type; 65| 6.24k| container_t *c2 = container_add(c, val & 0xFFFF, *type, &new_type); 66| 6.24k| *index = i; 67| 6.24k| if (c2 != c) { ------------------ | Branch (67:13): [True: 3, False: 6.24k] ------------------ 68| 3| container_free(c, *type); 69| 3| ra_set_container_at_index(ra, i, c2, new_type); 70| 3| *type = new_type; 71| 3| return c2; 72| 6.24k| } else { 73| 6.24k| return c; 74| 6.24k| } 75| 12.4k| } else { 76| 12.4k| array_container_t *new_ac = array_container_create(); 77| 12.4k| container_t *c = 78| 12.4k| container_add(new_ac, val & 0xFFFF, ARRAY_CONTAINER_TYPE, type); ------------------ | | 49| 12.4k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 79| | // we could just assume that it stays an array container 80| 12.4k| ra_insert_new_key_value_at(ra, -i - 1, hb, c, *type); 81| 12.4k| *index = -i - 1; 82| 12.4k| return c; 83| 12.4k| } 84| 18.7k|} roaring.c:add_bulk_impl: 106| 7.80M| uint32_t val) { 107| 7.80M| uint16_t key = val >> 16; 108| 7.80M| if (context->container == NULL || context->key != key) { ------------------ | Branch (108:9): [True: 0, False: 7.80M] | Branch (108:39): [True: 0, False: 7.80M] ------------------ 109| 0| uint8_t typecode; 110| 0| int idx; 111| 0| context->container = 112| 0| containerptr_roaring_bitmap_add(r, val, &typecode, &idx); 113| 0| context->typecode = typecode; 114| 0| context->idx = idx; 115| 0| context->key = key; 116| 7.80M| } else { 117| | // no need to seek the container, it is at hand 118| | // because we already have the container at hand, we can do the 119| | // insertion directly, bypassing the roaring_bitmap_add call 120| 7.80M| uint8_t new_typecode; 121| 7.80M| container_t *container2 = container_add( 122| 7.80M| context->container, val & 0xFFFF, context->typecode, &new_typecode); 123| 7.80M| if (container2 != context->container) { ------------------ | Branch (123:13): [True: 117, False: 7.80M] ------------------ 124| | // rare instance when we need to change the container type 125| 117| container_free(context->container, context->typecode); 126| 117| ra_set_container_at_index(&r->high_low_container, context->idx, 127| 117| container2, new_typecode); 128| 117| context->typecode = new_typecode; 129| 117| context->container = container2; 130| 117| } 131| 7.80M| } 132| 7.80M|} roaring.c:is_cow: 41| 105k|static inline bool is_cow(const roaring_bitmap_t *r) { 42| | return r->high_low_container.flags & ROARING_FLAG_COW; ------------------ | | 46| 105k|#define ROARING_FLAG_COW UINT8_C(0x1) ------------------ 43| 105k|} roaring.c:roaring_inplace_merge_bulk: 957| 2.27k| int left, int right, bool is_xor) { 958| 2.27k| roaring_array_t *ra1 = &x1->high_low_container; 959| 2.27k| const roaring_array_t *ra2 = &x2->high_low_container; 960| 2.27k| const bool cow2 = is_cow(x2); 961| 2.27k| const int length1 = ra1->size; 962| 2.27k| const int length2 = ra2->size; 963| | 964| 2.27k| int distinct = 0; 965| 2.27k| { 966| 2.27k| int l = left, r = right; 967| 57.5k| while (l < length1 && r < length2) { ------------------ | Branch (967:16): [True: 55.2k, False: 2.27k] | Branch (967:31): [True: 55.2k, False: 0] ------------------ 968| 55.2k| uint16_t k1 = ra1->keys[l]; 969| 55.2k| uint16_t k2 = ra2->keys[r]; 970| 55.2k| if (k1 < k2) { ------------------ | Branch (970:17): [True: 0, False: 55.2k] ------------------ 971| 0| l++; 972| 55.2k| } else if (k1 > k2) { ------------------ | Branch (972:24): [True: 102, False: 55.1k] ------------------ 973| 102| r++; 974| 55.1k| } else { 975| 55.1k| l++; 976| 55.1k| r++; 977| 55.1k| } 978| 55.2k| distinct++; 979| 55.2k| } 980| 2.27k| distinct += (length1 - l) + (length2 - r); 981| 2.27k| } 982| 2.27k| const int total = dst + distinct; 983| | 984| 2.27k| roaring_array_t merged; 985| 2.27k| ra_init_with_capacity(&merged, total > 0 ? (uint32_t)total : 1); ------------------ | Branch (985:36): [True: 2.20k, False: 78] ------------------ 986| | 987| 4.26k| for (int i = 0; i < dst; i++) { ------------------ | Branch (987:21): [True: 1.98k, False: 2.27k] ------------------ 988| 1.98k| ra_append(&merged, ra1->keys[i], ra1->containers[i], ra1->typecodes[i]); 989| 1.98k| } 990| | 991| 2.27k| uint8_t result_type = 0; 992| 57.5k| while (left < length1 && right < length2) { ------------------ | Branch (992:12): [True: 55.2k, False: 2.27k] | Branch (992:30): [True: 55.2k, False: 0] ------------------ 993| 55.2k| uint16_t k1 = ra1->keys[left]; 994| 55.2k| uint16_t k2 = ra2->keys[right]; 995| 55.2k| if (k1 < k2) { ------------------ | Branch (995:13): [True: 0, False: 55.2k] ------------------ 996| 0| ra_append(&merged, k1, ra1->containers[left], ra1->typecodes[left]); 997| 0| left++; 998| 55.2k| } else if (k1 > k2) { ------------------ | Branch (998:20): [True: 102, False: 55.1k] ------------------ 999| 102| uint8_t type2 = ra2->typecodes[right]; 1000| 102| container_t *c2 = 1001| 102| get_copy_of_container(ra2->containers[right], &type2, cow2); 1002| 102| if (cow2) { ------------------ | Branch (1002:17): [True: 0, False: 102] ------------------ 1003| 0| ra_set_container_at_index(ra2, right, c2, type2); 1004| 0| } 1005| 102| ra_append(&merged, k2, c2, type2); 1006| 102| right++; 1007| 55.1k| } else { 1008| 55.1k| uint8_t type1 = ra1->typecodes[left]; 1009| 55.1k| container_t *c1 = ra1->containers[left]; 1010| 55.1k| uint8_t type2 = ra2->typecodes[right]; 1011| 55.1k| container_t *c2 = ra2->containers[right]; 1012| 55.1k| if (is_xor) { ------------------ | Branch (1012:17): [True: 55.1k, False: 0] ------------------ 1013| 55.1k| container_t *c; 1014| 55.1k| if (type1 == SHARED_CONTAINER_TYPE) { ------------------ | | 51| 55.1k|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1014:21): [True: 0, False: 55.1k] ------------------ 1015| 0| c = container_xor(c1, type1, c2, type2, &result_type); 1016| 0| shared_container_free(CAST_shared(c1)); ------------------ | | 79| 0|#define CAST_shared(c) CAST(shared_container_t *, c) // safer downcast | | ------------------ | | | | 86| 0|#define CAST(type, value) ((type)value) | | ------------------ ------------------ 1017| 55.1k| } else { 1018| 55.1k| c = container_ixor(c1, type1, c2, type2, &result_type); 1019| 55.1k| } 1020| 55.1k| if (container_nonzero_cardinality(c, result_type)) { ------------------ | Branch (1020:21): [True: 0, False: 55.1k] ------------------ 1021| 0| ra_append(&merged, k1, c, result_type); 1022| 55.1k| } else { 1023| 55.1k| container_free(c, result_type); 1024| 55.1k| } 1025| 55.1k| } else { 1026| 0| if (container_is_full(c1, type1)) { ------------------ | Branch (1026:21): [True: 0, False: 0] ------------------ 1027| 0| ra_append(&merged, k1, c1, type1); 1028| 0| } else { 1029| 0| container_t *c = 1030| 0| (type1 == SHARED_CONTAINER_TYPE) ------------------ | | 51| 0|#define SHARED_CONTAINER_TYPE 4 ------------------ | Branch (1030:25): [True: 0, False: 0] ------------------ 1031| 0| ? container_or(c1, type1, c2, type2, &result_type) 1032| 0| : container_ior(c1, type1, c2, type2, &result_type); 1033| 0| if (c != c1) { ------------------ | Branch (1033:25): [True: 0, False: 0] ------------------ 1034| 0| container_free(c1, type1); 1035| 0| } 1036| 0| ra_append(&merged, k1, c, result_type); 1037| 0| } 1038| 0| } 1039| 55.1k| left++; 1040| 55.1k| right++; 1041| 55.1k| } 1042| 55.2k| } 1043| 2.27k| for (; left < length1; left++) { ------------------ | Branch (1043:12): [True: 0, False: 2.27k] ------------------ 1044| 0| ra_append(&merged, ra1->keys[left], ra1->containers[left], 1045| 0| ra1->typecodes[left]); 1046| 0| } 1047| 2.27k| for (; right < length2; right++) { ------------------ | Branch (1047:12): [True: 0, False: 2.27k] ------------------ 1048| 0| uint8_t type2 = ra2->typecodes[right]; 1049| 0| container_t *c2 = 1050| 0| get_copy_of_container(ra2->containers[right], &type2, cow2); 1051| 0| if (cow2) { ------------------ | Branch (1051:13): [True: 0, False: 0] ------------------ 1052| 0| ra_set_container_at_index(ra2, right, c2, type2); 1053| 0| } 1054| 0| ra_append(&merged, ra2->keys[right], c2, type2); 1055| 0| } 1056| | 1057| 2.27k| merged.flags = ra1->flags; 1058| 2.27k| ra_clear_without_containers(ra1); 1059| 2.27k| *ra1 = merged; 1060| 2.27k|} roaring.c:loadfirstvalue: 1809| 18.6k| roaring_uint32_iterator_t *newit) { 1810| 18.6k| if (iter_new_container_partial_init(newit)) { ------------------ | Branch (1810:9): [True: 12.3k, False: 6.27k] ------------------ 1811| 12.3k| uint16_t value = 0; 1812| 12.3k| newit->container_it = 1813| 12.3k| container_init_iterator(newit->container, newit->typecode, &value); 1814| 12.3k| newit->current_value = newit->highbits | value; 1815| 12.3k| } 1816| 18.6k| return newit->has_value; 1817| 18.6k|} roaring.c:iter_new_container_partial_init: 1779| 24.7k| roaring_uint32_iterator_t *newit) { 1780| 24.7k| newit->current_value = 0; 1781| 24.7k| if (newit->container_index >= newit->parent->high_low_container.size || ------------------ | Branch (1781:9): [True: 6.27k, False: 18.5k] ------------------ 1782| 18.5k| newit->container_index < 0) { ------------------ | Branch (1782:9): [True: 0, False: 18.5k] ------------------ 1783| 6.27k| newit->current_value = UINT32_MAX; 1784| 6.27k| return (newit->has_value = false); 1785| 6.27k| } 1786| 18.5k| newit->has_value = true; 1787| | // we precompute container, typecode and highbits so that successive 1788| | // iterators do not have to grab them from odd memory locations 1789| | // and have to worry about the (easily predicted) container_unwrap_shared 1790| | // call. 1791| 18.5k| newit->container = 1792| 18.5k| newit->parent->high_low_container.containers[newit->container_index]; 1793| 18.5k| newit->typecode = 1794| 18.5k| newit->parent->high_low_container.typecodes[newit->container_index]; 1795| 18.5k| newit->highbits = 1796| 18.5k| ((uint32_t) 1797| 18.5k| newit->parent->high_low_container.keys[newit->container_index]) 1798| 18.5k| << 16; 1799| 18.5k| newit->container = 1800| 18.5k| container_unwrap_shared(newit->container, &(newit->typecode)); 1801| | return true; 1802| 24.7k|} roaring.c:loadfirstvalue_largeorequal: 1840| 6.17k| roaring_uint32_iterator_t *newit, uint32_t val) { 1841| 6.17k| bool partial_init = iter_new_container_partial_init(newit); 1842| 6.17k| assert(partial_init); 1843| 6.17k| if (!partial_init) { ------------------ | Branch (1843:9): [True: 0, False: 6.17k] ------------------ 1844| 0| return false; 1845| 0| } 1846| 6.17k| uint16_t value = 0; 1847| 6.17k| newit->container_it = 1848| 6.17k| container_init_iterator(newit->container, newit->typecode, &value); 1849| 6.17k| bool found = container_iterator_lower_bound( 1850| 6.17k| newit->container, newit->typecode, &newit->container_it, &value, 1851| 6.17k| val & 0xFFFF); 1852| 6.17k| assert(found); 1853| 6.17k| if (!found) { ------------------ | Branch (1853:9): [True: 0, False: 6.17k] ------------------ 1854| 0| return false; 1855| 0| } 1856| 6.17k| newit->current_value = newit->highbits | value; 1857| | return true; 1858| 6.17k|} roaring.c:inplace_flip_container: 2214| 6.95k| uint16_t lb_start, uint16_t lb_end) { 2215| 6.95k| const int i = ra_get_index(x1_arr, hb); 2216| 6.95k| uint8_t ctype_in, ctype_out; 2217| 6.95k| container_t *flipped_container = NULL; 2218| 6.95k| if (i >= 0) { ------------------ | Branch (2218:9): [True: 6.45k, False: 504] ------------------ 2219| 6.45k| container_t *container_to_flip = 2220| 6.45k| ra_get_container_at_index(x1_arr, (uint16_t)i, &ctype_in); 2221| 6.45k| flipped_container = container_inot_range( 2222| 6.45k| container_to_flip, ctype_in, (uint32_t)lb_start, 2223| 6.45k| (uint32_t)(lb_end + 1), &ctype_out); 2224| | // if a new container was created, the old one was already freed 2225| 6.45k| if (container_nonzero_cardinality(flipped_container, ctype_out)) { ------------------ | Branch (2225:13): [True: 6.42k, False: 23] ------------------ 2226| 6.42k| ra_set_container_at_index(x1_arr, i, flipped_container, ctype_out); 2227| 6.42k| } else { 2228| 23| container_free(flipped_container, ctype_out); 2229| 23| ra_remove_at_index(x1_arr, i); 2230| 23| } 2231| | 2232| 6.45k| } else { 2233| 504| flipped_container = container_range_of_ones( 2234| 504| (uint32_t)lb_start, (uint32_t)(lb_end + 1), &ctype_out); 2235| 504| ra_insert_new_key_value_at(x1_arr, -i - 1, hb, flipped_container, 2236| 504| ctype_out); 2237| 504| } 2238| 6.95k|} roaring.c:inplace_fully_flip_container: 2265| 23.2k|static void inplace_fully_flip_container(roaring_array_t *x1_arr, uint16_t hb) { 2266| 23.2k| const int i = ra_get_index(x1_arr, hb); 2267| 23.2k| uint8_t ctype_in, ctype_out; 2268| 23.2k| container_t *flipped_container = NULL; 2269| 23.2k| if (i >= 0) { ------------------ | Branch (2269:9): [True: 6.84k, False: 16.3k] ------------------ 2270| 6.84k| container_t *container_to_flip = 2271| 6.84k| ra_get_container_at_index(x1_arr, (uint16_t)i, &ctype_in); 2272| 6.84k| flipped_container = 2273| 6.84k| container_inot(container_to_flip, ctype_in, &ctype_out); 2274| | 2275| 6.84k| if (container_nonzero_cardinality(flipped_container, ctype_out)) { ------------------ | Branch (2275:13): [True: 495, False: 6.35k] ------------------ 2276| 495| ra_set_container_at_index(x1_arr, i, flipped_container, ctype_out); 2277| 6.35k| } else { 2278| 6.35k| container_free(flipped_container, ctype_out); 2279| 6.35k| ra_remove_at_index(x1_arr, i); 2280| 6.35k| } 2281| | 2282| 16.3k| } else { 2283| 16.3k| flipped_container = container_range_of_ones(0U, 0x10000U, &ctype_out); 2284| 16.3k| ra_insert_new_key_value_at(x1_arr, -i - 1, hb, flipped_container, 2285| 16.3k| ctype_out); 2286| 16.3k| } 2287| 23.2k|} ra_init_with_capacity: 80| 64.7k|bool ra_init_with_capacity(roaring_array_t *new_ra, uint32_t cap) { 81| 64.7k| if (!new_ra) return false; ------------------ | Branch (81:9): [True: 0, False: 64.7k] ------------------ 82| 64.7k| ra_init(new_ra); 83| | 84| | // Containers hold 64Ki elements, so 64Ki containers is enough to hold 85| | // `0x10000 * 0x10000` (all 2^32) elements 86| 64.7k| if (cap > 0x10000) { ------------------ | Branch (86:9): [True: 0, False: 64.7k] ------------------ 87| 0| cap = 0x10000; 88| 0| } 89| | 90| 64.7k| if (cap > 0) { ------------------ | Branch (90:9): [True: 33.4k, False: 31.2k] ------------------ 91| 33.4k| void *bigalloc = roaring_malloc( 92| 33.4k| cap * (sizeof(uint16_t) + sizeof(container_t *) + sizeof(uint8_t))); 93| 33.4k| if (bigalloc == NULL) return false; ------------------ | Branch (93:13): [True: 0, False: 33.4k] ------------------ 94| 33.4k| new_ra->containers = (container_t **)bigalloc; 95| 33.4k| new_ra->keys = (uint16_t *)(new_ra->containers + cap); 96| 33.4k| new_ra->typecodes = (uint8_t *)(new_ra->keys + cap); 97| | // Narrowing is safe because of above check 98| 33.4k| new_ra->allocation_size = (int32_t)cap; 99| 33.4k| } 100| 64.7k| return true; 101| 64.7k|} ra_shrink_to_fit: 103| 68.7k|int ra_shrink_to_fit(roaring_array_t *ra) { 104| 68.7k| int savings = (ra->allocation_size - ra->size) * 105| 68.7k| (sizeof(uint16_t) + sizeof(container_t *) + sizeof(uint8_t)); 106| 68.7k| if (!realloc_array(ra, ra->size)) { ------------------ | Branch (106:9): [True: 0, False: 68.7k] ------------------ 107| 0| return 0; 108| 0| } 109| 68.7k| ra->allocation_size = ra->size; 110| 68.7k| return savings; 111| 68.7k|} ra_init: 113| 64.7k|void ra_init(roaring_array_t *new_ra) { 114| 64.7k| if (!new_ra) { ------------------ | Branch (114:9): [True: 0, False: 64.7k] ------------------ 115| 0| return; 116| 0| } 117| 64.7k| new_ra->keys = NULL; 118| 64.7k| new_ra->containers = NULL; 119| 64.7k| new_ra->typecodes = NULL; 120| | 121| 64.7k| new_ra->allocation_size = 0; 122| 64.7k| new_ra->size = 0; 123| 64.7k| new_ra->flags = 0; 124| 64.7k|} ra_overwrite: 127| 14.1k| bool copy_on_write) { 128| 14.1k| ra_clear_containers(dest); // we are going to overwrite them 129| 14.1k| if (source->size == 0) { // Note: can't call memcpy(NULL), even w/size ------------------ | Branch (129:9): [True: 20, False: 14.1k] ------------------ 130| 20| dest->size = 0; // <--- This is important. 131| 20| return true; // output was just cleared, so they match 132| 20| } 133| 14.1k| if (dest->allocation_size < source->size) { ------------------ | Branch (133:9): [True: 12.4k, False: 1.65k] ------------------ 134| 12.4k| if (!realloc_array(dest, source->size)) { ------------------ | Branch (134:13): [True: 0, False: 12.4k] ------------------ 135| 0| return false; 136| 0| } 137| 12.4k| } 138| 14.1k| dest->size = source->size; 139| 14.1k| memcpy(dest->keys, source->keys, dest->size * sizeof(uint16_t)); 140| | // we go through the containers, turning them into shared containers... 141| 14.1k| if (copy_on_write) { ------------------ | Branch (141:9): [True: 0, False: 14.1k] ------------------ 142| 0| for (int32_t i = 0; i < dest->size; ++i) { ------------------ | Branch (142:29): [True: 0, False: 0] ------------------ 143| 0| source->containers[i] = get_copy_of_container( 144| 0| source->containers[i], &source->typecodes[i], copy_on_write); 145| 0| } 146| | // we do a shallow copy to the other bitmap 147| 0| memcpy(dest->containers, source->containers, 148| 0| dest->size * sizeof(container_t *)); 149| 0| memcpy(dest->typecodes, source->typecodes, 150| 0| dest->size * sizeof(uint8_t)); 151| 14.1k| } else { 152| 14.1k| memcpy(dest->typecodes, source->typecodes, 153| 14.1k| dest->size * sizeof(uint8_t)); 154| 163k| for (int32_t i = 0; i < dest->size; i++) { ------------------ | Branch (154:29): [True: 149k, False: 14.1k] ------------------ 155| 149k| dest->containers[i] = 156| 149k| container_clone(source->containers[i], source->typecodes[i]); 157| 149k| if (dest->containers[i] == NULL) { ------------------ | Branch (157:17): [True: 0, False: 149k] ------------------ 158| 0| for (int32_t j = 0; j < i; j++) { ------------------ | Branch (158:37): [True: 0, False: 0] ------------------ 159| 0| container_free(dest->containers[j], dest->typecodes[j]); 160| 0| } 161| 0| ra_clear_without_containers(dest); 162| 0| return false; 163| 0| } 164| 149k| } 165| 14.1k| } 166| 14.1k| return true; 167| 14.1k|} ra_clear_containers: 169| 76.6k|void ra_clear_containers(roaring_array_t *ra) { 170| 516k| for (int32_t i = 0; i < ra->size; ++i) { ------------------ | Branch (170:25): [True: 440k, False: 76.6k] ------------------ 171| 440k| container_free(ra->containers[i], ra->typecodes[i]); 172| 440k| } 173| 76.6k|} ra_reset: 175| 62.4k|void ra_reset(roaring_array_t *ra) { 176| 62.4k| ra_clear_containers(ra); 177| 62.4k| ra->size = 0; 178| 62.4k| ra_shrink_to_fit(ra); 179| 62.4k|} ra_clear_without_containers: 181| 2.31k|void ra_clear_without_containers(roaring_array_t *ra) { 182| 2.31k| roaring_free( 183| 2.31k| ra->containers); // keys and typecodes are allocated with containers 184| 2.31k| ra->size = 0; 185| 2.31k| ra->allocation_size = 0; 186| 2.31k| ra->containers = NULL; 187| 2.31k| ra->keys = NULL; 188| | ra->typecodes = NULL; 189| 2.31k|} ra_clear: 191| 38|void ra_clear(roaring_array_t *ra) { 192| 38| ra_clear_containers(ra); 193| 38| ra_clear_without_containers(ra); 194| 38|} extend_array: 196| 85.1k|bool extend_array(roaring_array_t *ra, int32_t k) { 197| 85.1k| int32_t desired_size = ra->size + k; 198| 85.1k| const int32_t max_containers = 65536; 199| 85.1k| assert(desired_size <= max_containers); 200| 85.1k| if (desired_size > ra->allocation_size) { ------------------ | Branch (200:9): [True: 17.5k, False: 67.5k] ------------------ 201| 17.5k| int32_t new_capacity = 202| 17.5k| (ra->size < 1024) ? 2 * desired_size : 5 * desired_size / 4; ------------------ | Branch (202:13): [True: 17.5k, False: 0] ------------------ 203| 17.5k| if (new_capacity > max_containers) { ------------------ | Branch (203:13): [True: 0, False: 17.5k] ------------------ 204| 0| new_capacity = max_containers; 205| 0| } 206| | 207| 17.5k| return realloc_array(ra, new_capacity); 208| 17.5k| } 209| 67.5k| return true; 210| 85.1k|} ra_append: 213| 26.1k| uint8_t typecode) { 214| 26.1k| extend_array(ra, 1); 215| 26.1k| const int32_t pos = ra->size; 216| | 217| 26.1k| ra->keys[pos] = key; 218| 26.1k| ra->containers[pos] = c; 219| 26.1k| ra->typecodes[pos] = typecode; 220| 26.1k| ra->size++; 221| 26.1k|} ra_append_copy_range: 254| 25.1k| bool copy_on_write) { 255| 25.1k| extend_array(ra, end_index - start_index); 256| 196k| for (int32_t i = start_index; i < end_index; ++i) { ------------------ | Branch (256:35): [True: 171k, False: 25.1k] ------------------ 257| 171k| const int32_t pos = ra->size; 258| 171k| ra->keys[pos] = sa->keys[i]; 259| 171k| if (copy_on_write) { ------------------ | Branch (259:13): [True: 0, False: 171k] ------------------ 260| 0| sa->containers[i] = get_copy_of_container( 261| 0| sa->containers[i], &sa->typecodes[i], copy_on_write); 262| 0| ra->containers[pos] = sa->containers[i]; 263| 0| ra->typecodes[pos] = sa->typecodes[i]; 264| 171k| } else { 265| 171k| ra->containers[pos] = 266| 171k| container_clone(sa->containers[i], sa->typecodes[i]); 267| 171k| ra->typecodes[pos] = sa->typecodes[i]; 268| 171k| } 269| 171k| ra->size++; 270| 171k| } 271| 25.1k|} ra_advance_until_freeing: 340| 67|int32_t ra_advance_until_freeing(roaring_array_t *ra, uint16_t x, int32_t pos) { 341| 134| while (pos < ra->size && ra->keys[pos] < x) { ------------------ | Branch (341:12): [True: 134, False: 0] | Branch (341:30): [True: 67, False: 67] ------------------ 342| 67| container_free(ra->containers[pos], ra->typecodes[pos]); 343| 67| ++pos; 344| 67| } 345| 67| return pos; 346| 67|} ra_insert_new_key_value_at: 349| 32.5k| container_t *c, uint8_t typecode) { 350| 32.5k| extend_array(ra, 1); 351| | // May be an optimization opportunity with DIY memmove 352| 32.5k| memmove(&(ra->keys[i + 1]), &(ra->keys[i]), 353| 32.5k| sizeof(uint16_t) * (ra->size - i)); 354| 32.5k| memmove(&(ra->containers[i + 1]), &(ra->containers[i]), 355| 32.5k| sizeof(container_t *) * (ra->size - i)); 356| 32.5k| memmove(&(ra->typecodes[i + 1]), &(ra->typecodes[i]), 357| 32.5k| sizeof(uint8_t) * (ra->size - i)); 358| 32.5k| ra->keys[i] = key; 359| 32.5k| ra->containers[i] = c; 360| 32.5k| ra->typecodes[i] = typecode; 361| 32.5k| ra->size++; 362| 32.5k|} ra_downsize: 369| 12.0k|void ra_downsize(roaring_array_t *ra, int32_t new_length) { 370| | assert(new_length <= ra->size); 371| 12.0k| ra->size = new_length; 372| 12.0k|} ra_remove_at_index: 374| 6.43k|void ra_remove_at_index(roaring_array_t *ra, int32_t i) { 375| 6.43k| memmove(&(ra->containers[i]), &(ra->containers[i + 1]), 376| 6.43k| sizeof(container_t *) * (ra->size - i - 1)); 377| 6.43k| memmove(&(ra->keys[i]), &(ra->keys[i + 1]), 378| 6.43k| sizeof(uint16_t) * (ra->size - i - 1)); 379| 6.43k| memmove(&(ra->typecodes[i]), &(ra->typecodes[i + 1]), 380| 6.43k| sizeof(uint8_t) * (ra->size - i - 1)); 381| 6.43k| ra->size--; 382| 6.43k|} ra_remove_at_index_and_free: 384| 59|void ra_remove_at_index_and_free(roaring_array_t *ra, int32_t i) { 385| 59| container_free(ra->containers[i], ra->typecodes[i]); 386| 59| ra_remove_at_index(ra, i); 387| 59|} ra_copy_range: 395| 38| uint32_t new_begin) { 396| 38| assert(begin <= end); 397| 38| assert(new_begin < begin); 398| | 399| 38| const int range = end - begin; 400| | 401| | // We ensure to previously have freed overwritten containers 402| | // that are not copied elsewhere 403| | 404| 38| memmove(&(ra->containers[new_begin]), &(ra->containers[begin]), 405| 38| sizeof(container_t *) * range); 406| 38| memmove(&(ra->keys[new_begin]), &(ra->keys[begin]), 407| 38| sizeof(uint16_t) * range); 408| 38| memmove(&(ra->typecodes[new_begin]), &(ra->typecodes[begin]), 409| 38| sizeof(uint8_t) * range); 410| 38|} ra_shift_tail: 412| 1.85k|void ra_shift_tail(roaring_array_t *ra, int32_t count, int32_t distance) { 413| 1.85k| if (distance > 0) { ------------------ | Branch (413:9): [True: 1.24k, False: 615] ------------------ 414| 1.24k| extend_array(ra, distance); 415| 1.24k| } 416| 1.85k| int32_t srcpos = ra->size - count; 417| 1.85k| int32_t dstpos = srcpos + distance; 418| 1.85k| memmove(&(ra->keys[dstpos]), &(ra->keys[srcpos]), sizeof(uint16_t) * count); 419| 1.85k| memmove(&(ra->containers[dstpos]), &(ra->containers[srcpos]), 420| 1.85k| sizeof(container_t *) * count); 421| 1.85k| memmove(&(ra->typecodes[dstpos]), &(ra->typecodes[srcpos]), 422| 1.85k| sizeof(uint8_t) * count); 423| 1.85k| ra->size += distance; 424| 1.85k|} ra_to_uint32_array: 426| 6.24k|void ra_to_uint32_array(const roaring_array_t *ra, uint32_t *ans) { 427| 6.24k| size_t ctr = 0; 428| 69.5k| for (int32_t i = 0; i < ra->size; ++i) { ------------------ | Branch (428:25): [True: 63.2k, False: 6.24k] ------------------ 429| 63.2k| int num_added = container_to_uint32_array( 430| 63.2k| ans + ctr, ra->containers[i], ra->typecodes[i], 431| 63.2k| ((uint32_t)ra->keys[i]) << 16); 432| 63.2k| ctr += num_added; 433| 63.2k| } 434| 6.24k|} ra_has_run_container: 436| 18.7k|bool ra_has_run_container(const roaring_array_t *ra) { 437| 170k| for (int32_t k = 0; k < ra->size; ++k) { ------------------ | Branch (437:25): [True: 153k, False: 16.1k] ------------------ 438| 153k| if (get_container_type(ra->containers[k], ra->typecodes[k]) == ------------------ | Branch (438:13): [True: 2.54k, False: 151k] ------------------ 439| 153k| RUN_CONTAINER_TYPE) ------------------ | | 50| 153k|#define RUN_CONTAINER_TYPE 3 ------------------ 440| 2.54k| return true; 441| 153k| } 442| 16.1k| return false; 443| 18.7k|} ra_portable_header_size: 445| 12.4k|uint32_t ra_portable_header_size(const roaring_array_t *ra) { 446| 12.4k| if (ra_has_run_container(ra)) { ------------------ | Branch (446:9): [True: 2.54k, False: 9.94k] ------------------ 447| 2.54k| if (ra->size < ------------------ | Branch (447:13): [True: 2.54k, False: 0] ------------------ 448| 2.54k| NO_OFFSET_THRESHOLD) { // for small bitmaps, we omit the offsets 449| 2.54k| return 4 + (ra->size + 7) / 8 + 4 * ra->size; 450| 2.54k| } 451| 0| return 4 + (ra->size + 7) / 8 + 452| 0| 8 * ra->size; // - 4 because we pack the size with the cookie 453| 9.94k| } else { 454| 9.94k| return 4 + 4 + 8 * ra->size; 455| 9.94k| } 456| 12.4k|} ra_portable_size_in_bytes: 458| 12.4k|size_t ra_portable_size_in_bytes(const roaring_array_t *ra) { 459| 12.4k| size_t count = ra_portable_header_size(ra); 460| | 461| 92.5k| for (int32_t k = 0; k < ra->size; ++k) { ------------------ | Branch (461:25): [True: 80.0k, False: 12.4k] ------------------ 462| 80.0k| count += container_size_in_bytes(ra->containers[k], ra->typecodes[k]); 463| 80.0k| } 464| 12.4k| return count; 465| 12.4k|} ra_portable_serialize: 469| 6.24k|size_t ra_portable_serialize(const roaring_array_t *ra, char *buf) { 470| 6.24k| char *initbuf = buf; 471| 6.24k| uint32_t startOffset = 0; 472| 6.24k| bool hasrun = ra_has_run_container(ra); 473| 6.24k| if (hasrun) { ------------------ | Branch (473:9): [True: 0, False: 6.24k] ------------------ 474| 0| uint32_t cookie = SERIAL_COOKIE | ((uint32_t)(ra->size - 1) << 16); 475| 0| uint32_t cookie_le = croaring_htole32(cookie); ------------------ | | 518| 0|#define croaring_htole32(x) (x) ------------------ 476| 0| memcpy(buf, &cookie_le, sizeof(cookie_le)); 477| 0| buf += sizeof(cookie_le); 478| 0| uint32_t s = (ra->size + 7) / 8; 479| 0| memset(buf, 0, s); 480| 0| for (int32_t i = 0; i < ra->size; ++i) { ------------------ | Branch (480:29): [True: 0, False: 0] ------------------ 481| 0| if (get_container_type(ra->containers[i], ra->typecodes[i]) == ------------------ | Branch (481:17): [True: 0, False: 0] ------------------ 482| 0| RUN_CONTAINER_TYPE) { ------------------ | | 50| 0|#define RUN_CONTAINER_TYPE 3 ------------------ 483| 0| buf[i / 8] |= 1 << (i % 8); 484| 0| } 485| 0| } 486| 0| buf += s; 487| 0| if (ra->size < NO_OFFSET_THRESHOLD) { ------------------ | Branch (487:13): [True: 0, False: 0] ------------------ 488| 0| startOffset = 4 + 4 * ra->size + s; 489| 0| } else { 490| 0| startOffset = 4 + 8 * ra->size + s; 491| 0| } 492| 6.24k| } else { // backwards compatibility 493| 6.24k| uint32_t cookie = SERIAL_COOKIE_NO_RUNCONTAINER; 494| 6.24k| uint32_t cookie_le = croaring_htole32(cookie); ------------------ | | 518| 6.24k|#define croaring_htole32(x) (x) ------------------ 495| 6.24k| memcpy(buf, &cookie_le, sizeof(cookie_le)); 496| 6.24k| buf += sizeof(cookie_le); 497| 6.24k| uint32_t size_le = croaring_htole32((uint32_t)ra->size); ------------------ | | 518| 6.24k|#define croaring_htole32(x) (x) ------------------ 498| 6.24k| memcpy(buf, &size_le, sizeof(size_le)); 499| 6.24k| buf += sizeof(size_le); 500| | 501| 6.24k| startOffset = 4 + 4 + 4 * ra->size + 4 * ra->size; 502| 6.24k| } 503| 80.0k| for (int32_t k = 0; k < ra->size; ++k) { ------------------ | Branch (503:25): [True: 73.8k, False: 6.24k] ------------------ 504| 73.8k| uint16_t key_le = croaring_htole16(ra->keys[k]); ------------------ | | 517| 73.8k|#define croaring_htole16(x) (x) ------------------ 505| 73.8k| memcpy(buf, &key_le, sizeof(key_le)); 506| 73.8k| buf += sizeof(key_le); 507| | // get_cardinality returns a value in [1,1<<16], subtracting one 508| | // we get [0,1<<16 - 1] which fits in 16 bits 509| 73.8k| uint16_t card = (uint16_t)(container_get_cardinality(ra->containers[k], 510| 73.8k| ra->typecodes[k]) - 511| 73.8k| 1); 512| 73.8k| uint16_t card_le = croaring_htole16(card); ------------------ | | 517| 73.8k|#define croaring_htole16(x) (x) ------------------ 513| 73.8k| memcpy(buf, &card_le, sizeof(card_le)); 514| 73.8k| buf += sizeof(card_le); 515| 73.8k| } 516| 6.24k| if ((!hasrun) || (ra->size >= NO_OFFSET_THRESHOLD)) { ------------------ | Branch (516:9): [True: 6.24k, False: 0] | Branch (516:22): [True: 0, False: 0] ------------------ 517| | // writing the containers offsets 518| 80.0k| for (int32_t k = 0; k < ra->size; k++) { ------------------ | Branch (518:29): [True: 73.8k, False: 6.24k] ------------------ 519| 73.8k| uint32_t off_le = croaring_htole32(startOffset); ------------------ | | 518| 73.8k|#define croaring_htole32(x) (x) ------------------ 520| 73.8k| memcpy(buf, &off_le, sizeof(off_le)); 521| 73.8k| buf += sizeof(off_le); 522| 73.8k| startOffset = 523| 73.8k| startOffset + 524| 73.8k| container_size_in_bytes(ra->containers[k], ra->typecodes[k]); 525| 73.8k| } 526| 6.24k| } 527| 80.0k| for (int32_t k = 0; k < ra->size; ++k) { ------------------ | Branch (527:25): [True: 73.8k, False: 6.24k] ------------------ 528| 73.8k| buf += container_write(ra->containers[k], ra->typecodes[k], buf); 529| 73.8k| } 530| 6.24k| return buf - initbuf; 531| 6.24k|} ra_portable_deserialize: 634| 12.4k| const size_t maxbytes, size_t *readbytes) { 635| 12.4k| *readbytes = sizeof(int32_t); // for cookie 636| 12.4k| if (*readbytes > maxbytes) { ------------------ | Branch (636:9): [True: 5.86k, False: 6.62k] ------------------ 637| | // Ran out of bytes while reading first 4 bytes. 638| 5.86k| return false; 639| 5.86k| } 640| 6.62k| uint32_t cookie; 641| 6.62k| memcpy(&cookie, buf, sizeof(int32_t)); 642| 6.62k| cookie = croaring_letoh32(cookie); ------------------ | | 524| 6.62k|#define croaring_letoh32(x) croaring_htole32(x) | | ------------------ | | | | 518| 6.62k|#define croaring_htole32(x) (x) | | ------------------ ------------------ 643| 6.62k| buf += sizeof(uint32_t); 644| 6.62k| if ((cookie & 0xFFFF) != SERIAL_COOKIE && ------------------ | Branch (644:9): [True: 6.55k, False: 76] ------------------ 645| 6.55k| cookie != SERIAL_COOKIE_NO_RUNCONTAINER) { ------------------ | Branch (645:9): [True: 241, False: 6.31k] ------------------ 646| | // "I failed to find one of the right cookies. 647| 241| return false; 648| 241| } 649| 6.38k| int32_t size; 650| | 651| 6.38k| if ((cookie & 0xFFFF) == SERIAL_COOKIE) ------------------ | Branch (651:9): [True: 76, False: 6.31k] ------------------ 652| 76| size = (cookie >> 16) + 1; 653| 6.31k| else { 654| 6.31k| *readbytes += sizeof(int32_t); 655| 6.31k| if (*readbytes > maxbytes) { ------------------ | Branch (655:13): [True: 2, False: 6.30k] ------------------ 656| | // Ran out of bytes while reading second part of the cookie. 657| 2| return false; 658| 2| } 659| 6.30k| uint32_t size_le; 660| 6.30k| memcpy(&size_le, buf, sizeof(int32_t)); 661| 6.30k| size = (int32_t)croaring_letoh32(size_le); ------------------ | | 524| 6.30k|#define croaring_letoh32(x) croaring_htole32(x) | | ------------------ | | | | 518| 6.30k|#define croaring_htole32(x) (x) | | ------------------ ------------------ 662| 6.30k| buf += sizeof(uint32_t); 663| 6.30k| } 664| 6.38k| if (size < 0) { ------------------ | Branch (664:9): [True: 18, False: 6.36k] ------------------ 665| | // You cannot have a negative number of containers, the data must be 666| | // corrupted. 667| 18| return false; 668| 18| } 669| 6.36k| if (size > (1 << 16)) { ------------------ | Branch (669:9): [True: 16, False: 6.35k] ------------------ 670| | // You cannot have so many containers, the data must be corrupted. 671| 16| return false; 672| 16| } 673| 6.35k| const char *bitmapOfRunContainers = NULL; 674| 6.35k| bool hasrun = (cookie & 0xFFFF) == SERIAL_COOKIE; 675| 6.35k| if (hasrun) { ------------------ | Branch (675:9): [True: 76, False: 6.27k] ------------------ 676| 76| int32_t s = (size + 7) / 8; 677| 76| *readbytes += s; 678| 76| if (*readbytes > maxbytes) { // data is corrupted? ------------------ | Branch (678:13): [True: 22, False: 54] ------------------ 679| | // Ran out of bytes while reading run bitmap. 680| 22| return false; 681| 22| } 682| 54| bitmapOfRunContainers = buf; 683| 54| buf += s; 684| 54| } 685| 6.32k| const char *keyscards = buf; 686| | 687| 6.32k| *readbytes += size * 2 * sizeof(uint16_t); 688| 6.32k| if (*readbytes > maxbytes) { ------------------ | Branch (688:9): [True: 32, False: 6.29k] ------------------ 689| | // Ran out of bytes while reading key-cardinality array. 690| 32| return false; 691| 32| } 692| 6.29k| buf += size * 2 * sizeof(uint16_t); 693| | 694| 6.29k| bool is_ok = ra_init_with_capacity(answer, size); 695| 6.29k| if (!is_ok) { ------------------ | Branch (695:9): [True: 0, False: 6.29k] ------------------ 696| | // Failed to allocate memory for roaring array. Bailing out. 697| 0| return false; 698| 0| } 699| | 700| 80.3k| for (int32_t k = 0; k < size; ++k) { ------------------ | Branch (700:25): [True: 74.0k, False: 6.29k] ------------------ 701| 74.0k| uint16_t tmp; 702| 74.0k| memcpy(&tmp, keyscards + 4 * k, sizeof(tmp)); 703| 74.0k| answer->keys[k] = croaring_letoh16(tmp); ------------------ | | 523| 74.0k|#define croaring_letoh16(x) croaring_htole16(x) | | ------------------ | | | | 517| 74.0k|#define croaring_htole16(x) (x) | | ------------------ ------------------ 704| 74.0k| } 705| 6.29k| if ((!hasrun) || (size >= NO_OFFSET_THRESHOLD)) { ------------------ | Branch (705:9): [True: 6.25k, False: 45] | Branch (705:22): [True: 21, False: 24] ------------------ 706| 6.27k| *readbytes += size * 4; 707| 6.27k| if (*readbytes > maxbytes) { // data is corrupted? ------------------ | Branch (707:13): [True: 9, False: 6.26k] ------------------ 708| | // Ran out of bytes while reading offsets. 709| 9| ra_clear(answer); // we need to clear the containers already 710| | // allocated, and the roaring array 711| 9| return false; 712| 9| } 713| | 714| | // skipping the offsets 715| 6.26k| buf += size * 4; 716| 6.26k| } 717| | // Reading the containers 718| 80.1k| for (int32_t k = 0; k < size; ++k) { ------------------ | Branch (718:25): [True: 73.9k, False: 6.25k] ------------------ 719| 73.9k| uint16_t tmp; 720| 73.9k| memcpy(&tmp, keyscards + 4 * k + 2, sizeof(tmp)); 721| 73.9k| tmp = croaring_letoh16(tmp); ------------------ | | 523| 73.9k|#define croaring_letoh16(x) croaring_htole16(x) | | ------------------ | | | | 517| 73.9k|#define croaring_htole16(x) (x) | | ------------------ ------------------ 722| 73.9k| uint32_t thiscard = tmp + 1; 723| 73.9k| bool isbitmap = (thiscard > DEFAULT_MAX_SIZE); 724| 73.9k| bool isrun = false; 725| 73.9k| if (hasrun) { ------------------ | Branch (725:13): [True: 123, False: 73.8k] ------------------ 726| 123| if ((bitmapOfRunContainers[k / 8] & (1 << (k % 8))) != 0) { ------------------ | Branch (726:17): [True: 64, False: 59] ------------------ 727| 64| isbitmap = false; 728| 64| isrun = true; 729| 64| } 730| 123| } 731| 73.9k| if (isbitmap) { ------------------ | Branch (731:13): [True: 66.9k, False: 6.93k] ------------------ 732| | // we check that the read is allowed 733| 66.9k| size_t containersize = 734| 66.9k| BITSET_CONTAINER_SIZE_IN_WORDS * sizeof(uint64_t); 735| 66.9k| *readbytes += containersize; 736| 66.9k| if (*readbytes > maxbytes) { ------------------ | Branch (736:17): [True: 7, False: 66.9k] ------------------ 737| | // Running out of bytes while reading a bitset container. 738| 7| ra_clear(answer); // we need to clear the containers already 739| | // allocated, and the roaring array 740| 7| return false; 741| 7| } 742| | // it is now safe to read 743| 66.9k| bitset_container_t *c = bitset_container_create_uninitialized(); 744| 66.9k| if (c == NULL) { // memory allocation failure ------------------ | Branch (744:17): [True: 0, False: 66.9k] ------------------ 745| | // Failed to allocate memory for a bitset container. 746| 0| ra_clear(answer); // we need to clear the containers already 747| | // allocated, and the roaring array 748| 0| return false; 749| 0| } 750| 66.9k| answer->size++; 751| 66.9k| buf += bitset_container_read(thiscard, c, buf); 752| 66.9k| answer->containers[k] = c; 753| 66.9k| answer->typecodes[k] = BITSET_CONTAINER_TYPE; ------------------ | | 48| 66.9k|#define BITSET_CONTAINER_TYPE 1 ------------------ 754| 66.9k| } else if (isrun) { ------------------ | Branch (754:20): [True: 64, False: 6.87k] ------------------ 755| | // we check that the read is allowed 756| 64| *readbytes += sizeof(uint16_t); 757| 64| if (*readbytes > maxbytes) { ------------------ | Branch (757:17): [True: 9, False: 55] ------------------ 758| | // Running out of bytes while reading a run container (header). 759| 9| ra_clear(answer); // we need to clear the containers already 760| | // allocated, and the roaring array 761| 9| return false; 762| 9| } 763| 55| uint16_t n_runs; 764| 55| memcpy(&n_runs, buf, sizeof(uint16_t)); 765| 55| n_runs = croaring_letoh16(n_runs); ------------------ | | 523| 55|#define croaring_letoh16(x) croaring_htole16(x) | | ------------------ | | | | 517| 55|#define croaring_htole16(x) (x) | | ------------------ ------------------ 766| 55| size_t containersize = n_runs * sizeof(rle16_t); 767| 55| *readbytes += containersize; 768| 55| if (*readbytes > maxbytes) { // data is corrupted? ------------------ | Branch (768:17): [True: 6, False: 49] ------------------ 769| | // Running out of bytes while reading a run container. 770| 6| ra_clear(answer); // we need to clear the containers already 771| | // allocated, and the roaring array 772| 6| return false; 773| 6| } 774| | // it is now safe to read 775| | 776| 49| run_container_t *c = run_container_create(); 777| 49| if (c == NULL) { // memory allocation failure ------------------ | Branch (777:17): [True: 0, False: 49] ------------------ 778| | // Failed to allocate memory for a run container. 779| 0| ra_clear(answer); // we need to clear the containers already 780| | // allocated, and the roaring array 781| 0| return false; 782| 0| } 783| 49| answer->size++; 784| 49| buf += run_container_read(thiscard, c, buf); 785| 49| answer->containers[k] = c; 786| 49| answer->typecodes[k] = RUN_CONTAINER_TYPE; ------------------ | | 50| 49|#define RUN_CONTAINER_TYPE 3 ------------------ 787| 6.87k| } else { 788| | // we check that the read is allowed 789| 6.87k| size_t containersize = thiscard * sizeof(uint16_t); 790| 6.87k| *readbytes += containersize; 791| 6.87k| if (*readbytes > maxbytes) { // data is corrupted? ------------------ | Branch (791:17): [True: 7, False: 6.86k] ------------------ 792| | // Running out of bytes while reading an array container. 793| 7| ra_clear(answer); // we need to clear the containers already 794| | // allocated, and the roaring array 795| 7| return false; 796| 7| } 797| | // it is now safe to read 798| 6.86k| array_container_t *c = 799| 6.86k| array_container_create_given_capacity(thiscard); 800| 6.86k| if (c == NULL) { // memory allocation failure ------------------ | Branch (800:17): [True: 0, False: 6.86k] ------------------ 801| | // Failed to allocate memory for an array container. 802| 0| ra_clear(answer); // we need to clear the containers already 803| | // allocated, and the roaring array 804| 0| return false; 805| 0| } 806| 6.86k| answer->size++; 807| 6.86k| buf += array_container_read(thiscard, c, buf); 808| 6.86k| answer->containers[k] = c; 809| 6.86k| answer->typecodes[k] = ARRAY_CONTAINER_TYPE; ------------------ | | 49| 6.86k|#define ARRAY_CONTAINER_TYPE 2 ------------------ 810| 6.86k| } 811| 73.9k| } 812| 6.25k| return true; 813| 6.28k|} roaring_array.c:realloc_array: 42| 98.7k|static bool realloc_array(roaring_array_t *ra, int32_t new_capacity) { 43| | // 44| | // Note: not implemented using C's realloc(), because the memory layout is 45| | // Struct-of-Arrays vs. Array-of-Structs: 46| | // https://github.com/RoaringBitmap/CRoaring/issues/256 47| | 48| 98.7k| if (new_capacity == 0) { ------------------ | Branch (48:9): [True: 62.4k, False: 36.2k] ------------------ 49| 62.4k| roaring_free(ra->containers); 50| 62.4k| ra->containers = NULL; 51| 62.4k| ra->keys = NULL; 52| 62.4k| ra->typecodes = NULL; 53| 62.4k| ra->allocation_size = 0; 54| 62.4k| return true; 55| 62.4k| } 56| 36.2k| const size_t memoryneeded = 57| 36.2k| new_capacity * 58| 36.2k| (sizeof(uint16_t) + sizeof(container_t *) + sizeof(uint8_t)); 59| 36.2k| void *bigalloc = roaring_malloc(memoryneeded); 60| 36.2k| if (!bigalloc) return false; ------------------ | Branch (60:9): [True: 0, False: 36.2k] ------------------ 61| 36.2k| void *oldbigalloc = ra->containers; 62| 36.2k| container_t **newcontainers = (container_t **)bigalloc; 63| 36.2k| uint16_t *newkeys = (uint16_t *)(newcontainers + new_capacity); 64| 36.2k| uint8_t *newtypecodes = (uint8_t *)(newkeys + new_capacity); 65| 36.2k| assert((char *)(newtypecodes + new_capacity) == 66| 36.2k| (char *)bigalloc + memoryneeded); 67| 36.2k| if (ra->size > 0) { ------------------ | Branch (67:9): [True: 11.3k, False: 24.9k] ------------------ 68| 11.3k| memcpy(newcontainers, ra->containers, sizeof(container_t *) * ra->size); 69| 11.3k| memcpy(newkeys, ra->keys, sizeof(uint16_t) * ra->size); 70| 11.3k| memcpy(newtypecodes, ra->typecodes, sizeof(uint8_t) * ra->size); 71| 11.3k| } 72| 36.2k| ra->containers = newcontainers; 73| 36.2k| ra->keys = newkeys; 74| 36.2k| ra->typecodes = newtypecodes; 75| 36.2k| ra->allocation_size = new_capacity; 76| 36.2k| roaring_free(oldbigalloc); 77| | return true; 78| 36.2k|} _Z17ConsumeVecInRangeR18FuzzedDataProvidermjj: 24| 12.4k| uint32_t max_value) { 25| 12.4k| std::vector result = {0}; 26| 12.4k| result.resize(length); 27| 12.4k| std::generate(result.begin(), result.end(), [&]() { 28| 12.4k| return fdp.ConsumeIntegralInRange(min_value, max_value); 29| 12.4k| }); 30| 12.4k| return result; 31| 12.4k|} LLVMFuzzerTestOneInput: 33| 6.24k|extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { 34| | /** 35| | * A bitmap may contain up to 2**32 elements. Later this function will 36| | * output the content to an array where each element uses 32 bits of 37| | * storage. That would use 16 GB. Thus this function is bound to run out of 38| | * memory. 39| | * 40| | * Even without the full serialization to a 32-bit array, a bitmap may still 41| | * use over 512 MB in the normal course of operation: that is to be expected 42| | * since it can represent all sets of integers in [0,2**32]. This function 43| | * may hold several bitmaps in memory at once, so it can require gigabytes 44| | * of memory (without bugs). Hence, unless it has a generous memory 45| | * capacity, this function will run out of memory almost certainly. 46| | * 47| | * For sanity, we may limit the range to, say, 10,000,000 which will use 38 48| | * MB or so. With such a limited range, if we run out of memory, then we can 49| | * almost certain that it has to do with a genuine bug. 50| | */ 51| | 52| 6.24k| uint32_t range_start = 0; 53| 6.24k| uint32_t range_end = 10'000'000; 54| | 55| | /** 56| | * We are not solely dependent on the range [range_start, range_end) because 57| | * ConsumeVecInRange below produce integers in a small range starting at 0. 58| | */ 59| | 60| 6.24k| FuzzedDataProvider fdp(data, size); 61| | /** 62| | * The next line was ConsumeVecInRange(fdp, 500, 0, 1000) but it would pick 63| | * 500 values at random from 0, 1000, making almost certain that all of the 64| | * values are picked. It seems more useful to pick 500 values in the range 65| | * 0,1000. 66| | */ 67| 6.24k| std::vector bitmap_data_a = ConsumeVecInRange(fdp, 500, 0, 1000); 68| 6.24k| roaring::Roaring a(bitmap_data_a.size(), bitmap_data_a.data()); 69| 6.24k| a.runOptimize(); 70| 6.24k| a.shrinkToFit(); 71| | 72| 6.24k| std::vector bitmap_data_b = ConsumeVecInRange(fdp, 500, 0, 1000); 73| 6.24k| roaring::Roaring b(bitmap_data_b.size(), bitmap_data_b.data()); 74| 6.24k| b.runOptimize(); 75| 6.24k| b.add(fdp.ConsumeIntegralInRange(range_start, range_end)); 76| 6.24k| b.addChecked(fdp.ConsumeIntegralInRange(range_start, range_end)); 77| 6.24k| b.addRange(fdp.ConsumeIntegralInRange(range_start, range_end), 78| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 79| | // add half of a to b. 80| 6.24k| b.addMany(bitmap_data_a.size() / 2, bitmap_data_a.data()); 81| 6.24k| b.remove(fdp.ConsumeIntegralInRange(range_start, range_end)); 82| 6.24k| b.removeChecked( 83| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 84| 6.24k| b.removeRange(fdp.ConsumeIntegralInRange(range_start, range_end), 85| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 86| 6.24k| b.removeRangeClosed( 87| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end), 88| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 89| 6.24k| b.maximum(); 90| 6.24k| b.minimum(); 91| 6.24k| b.contains(fdp.ConsumeIntegralInRange(range_start, range_end)); 92| 6.24k| b.containsRange( 93| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end), 94| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 95| | 96| 6.24k| uint32_t element = 0; 97| 6.24k| a.select(fdp.ConsumeIntegralInRange(0, 1000), &element); 98| 6.24k| a.intersect(b); 99| 6.24k| a.jaccard_index(b); 100| 6.24k| a.or_cardinality(b); 101| 6.24k| a.andnot_cardinality(b); 102| 6.24k| a.xor_cardinality(b); 103| 6.24k| a.rank(fdp.ConsumeIntegralInRange(0, 5000)); 104| 6.24k| a.getSizeInBytes(); 105| | 106| 6.24k| roaring::Roaring c = a & b; 107| 6.24k| roaring::Roaring d = a - b; 108| 6.24k| roaring::Roaring e = a | b; 109| 6.24k| roaring::Roaring f = a ^ b; 110| 6.24k| a |= e; 111| 6.24k| a &= b; 112| 6.24k| a -= c; 113| 6.24k| a ^= f; 114| | 115| 6.24k| volatile bool is_equal = (a == b); 116| | 117| 6.24k| std::vector b_as_array = {0}; 118| 6.24k| b_as_array.resize(b.cardinality()); 119| 6.24k| b.isEmpty(); 120| 6.24k| b.toUint32Array(b_as_array.data()); 121| | 122| 6.24k| a.isSubset(b); 123| 6.24k| a.isStrictSubset(b); 124| 6.24k| b.flip(fdp.ConsumeIntegralInRange(range_start, range_end), 125| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 126| 6.24k| b.flipClosed(fdp.ConsumeIntegralInRange(range_start, range_end), 127| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 128| 6.24k| b.removeRunCompression(); 129| | 130| | // Move/copy constructors 131| 6.24k| roaring::Roaring copied = b; 132| 6.24k| roaring::Roaring moved = std::move(b); 133| | 134| | // Asignment operators 135| 6.24k| b = copied; 136| 6.24k| b = std::move(moved); 137| | 138| | // Safe read from serialized 139| 6.24k| std::vector read_buffer = fdp.ConsumeBytes(100); 140| 6.24k| try { 141| 6.24k| roaring::Roaring read_safely = 142| 6.24k| roaring::Roaring::readSafe(read_buffer.data(), read_buffer.size()); 143| | // The above is guaranteed to be safe. However, read_safely is maybe 144| | // in an improper state and it cannot be used safely (including for 145| | // reserialization). 146| 6.24k| } catch (...) { 147| 6.23k| } 148| | 149| | // The bitmap b can be serialized and re-read. 150| 6.24k| std::size_t expected_size_in_bytes = b.getSizeInBytes(); 151| 6.24k| std::vector buffer(expected_size_in_bytes); 152| 6.24k| std::size_t size_in_bytes = b.write(buffer.data()); 153| 6.24k| assert(expected_size_in_bytes == size_in_bytes); ------------------ | Branch (153:5): [True: 6.24k, False: 0] ------------------ 154| 6.24k| roaring::Roaring bread = 155| 6.24k| roaring::Roaring::readSafe(buffer.data(), size_in_bytes); 156| 6.24k| assert(bread == b); ------------------ | Branch (156:5): [True: 6.24k, False: 0] ------------------ 157| | 158| 6.24k| f.toString(); 159| | 160| 6.24k| volatile int unused = 0; 161| | 162| 335k| for (roaring::Roaring::const_iterator i = a.begin(); i != a.end(); i++) { ------------------ | Branch (162:58): [True: 329k, False: 6.24k] ------------------ 163| 329k| unused++; 164| 329k| } 165| | 166| 6.24k| roaring::Roaring::const_iterator b_iter = b.begin(); 167| 6.24k| b_iter.equalorlarger( 168| 6.24k| fdp.ConsumeIntegralInRange(range_start, range_end)); 169| | 170| 6.24k| return 0; 171| 6.24k|} croaring_fuzzer_cc.cc:_ZZ17ConsumeVecInRangeR18FuzzedDataProvidermjjENK3$_0clEv: 27| 6.24M| std::generate(result.begin(), result.end(), [&]() { 28| 6.24M| return fdp.ConsumeIntegralInRange(min_value, max_value); 29| 6.24M| });