Coverage Report

Created: 2026-06-08 06:48

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/gmp/mpn/toom22_mul.c
Line
Count
Source
1
/* mpn_toom22_mul -- Multiply {ap,an} and {bp,bn} where an >= bn.  Or more
2
   accurately, bn <= an < 2bn.
3
4
   Contributed to the GNU project by Torbjorn Granlund.
5
6
   THE FUNCTION IN THIS FILE IS INTERNAL WITH A MUTABLE INTERFACE.  IT IS ONLY
7
   SAFE TO REACH IT THROUGH DOCUMENTED INTERFACES.  IN FACT, IT IS ALMOST
8
   GUARANTEED THAT IT WILL CHANGE OR DISAPPEAR IN A FUTURE GNU MP RELEASE.
9
10
Copyright 2006-2010, 2012, 2014, 2018, 2020 Free Software Foundation, Inc.
11
12
This file is part of the GNU MP Library.
13
14
The GNU MP Library is free software; you can redistribute it and/or modify
15
it under the terms of either:
16
17
  * the GNU Lesser General Public License as published by the Free
18
    Software Foundation; either version 3 of the License, or (at your
19
    option) any later version.
20
21
or
22
23
  * the GNU General Public License as published by the Free Software
24
    Foundation; either version 2 of the License, or (at your option) any
25
    later version.
26
27
or both in parallel, as here.
28
29
The GNU MP Library is distributed in the hope that it will be useful, but
30
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
31
or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
32
for more details.
33
34
You should have received copies of the GNU General Public License and the
35
GNU Lesser General Public License along with the GNU MP Library.  If not,
36
see https://www.gnu.org/licenses/.  */
37
38
39
#include "gmp-impl.h"
40
41
/* Evaluate in: -1, 0, +inf
42
43
  <-s--><--n-->
44
   ____ ______
45
  |_a1_|___a0_|
46
   |b1_|___b0_|
47
   <-t-><--n-->
48
49
  v0  =  a0     * b0       #   A(0)*B(0)
50
  vm1 = (a0- a1)*(b0- b1)  #  A(-1)*B(-1)
51
  vinf=      a1 *     b1   # A(inf)*B(inf)
52
*/
53
54
#if TUNE_PROGRAM_BUILD || WANT_FAT_BINARY
55
#define MAYBE_mul_toom22   1
56
#else
57
#define MAYBE_mul_toom22            \
58
204M
  (MUL_TOOM33_THRESHOLD >= 2 * MUL_TOOM22_THRESHOLD)
59
#endif
60
61
#define TOOM22_MUL_N_REC(p, a, b, n, ws)        \
62
102M
  do {                 \
63
102M
    if (! MAYBE_mul_toom22            \
64
102M
  || BELOW_THRESHOLD (n, MUL_TOOM22_THRESHOLD))     \
65
102M
      mpn_mul_basecase (p, a, n, b, n);         \
66
102M
    else                \
67
102M
      mpn_toom22_mul (p, a, n, b, n, ws);       \
68
102M
  } while (0)
69
70
/* Normally, this calls mul_basecase or toom22_mul.  But when when the fraction
71
   MUL_TOOM33_THRESHOLD / MUL_TOOM22_THRESHOLD is large, an initially small
72
   relative unbalance will become a larger and larger relative unbalance with
73
   each recursion (the difference s-t will be invariant over recursive calls).
74
   Therefore, we need to call toom32_mul.  FIXME: Suppress depending on
75
   MUL_TOOM33_THRESHOLD / MUL_TOOM22_THRESHOLD and on MUL_TOOM22_THRESHOLD.  */
76
#define TOOM22_MUL_REC(p, a, an, b, bn, ws)       \
77
6.33k
  do {                 \
78
6.33k
    if (! MAYBE_mul_toom22            \
79
6.33k
  || BELOW_THRESHOLD (bn, MUL_TOOM22_THRESHOLD))      \
80
6.33k
      mpn_mul_basecase (p, a, an, b, bn);       \
81
6.33k
    else if (4 * an < 5 * bn)           \
82
512
      mpn_toom22_mul (p, a, an, b, bn, ws);       \
83
512
    else                \
84
512
      mpn_toom32_mul (p, a, an, b, bn, ws);       \
85
6.33k
  } while (0)
86
87
void
88
mpn_toom22_mul (mp_ptr pp,
89
    mp_srcptr ap, mp_size_t an,
90
    mp_srcptr bp, mp_size_t bn,
91
    mp_ptr scratch)
92
34.1M
{
93
34.1M
  const int __gmpn_cpuvec_initialized = 1;
94
34.1M
  mp_size_t n, s, t;
95
34.1M
  int vm1_neg;
96
34.1M
  mp_limb_t cy, cy2;
97
34.1M
  mp_ptr asm1;
98
34.1M
  mp_ptr bsm1;
99
100
81.8M
#define a0  ap
101
55.6M
#define a1  (ap + n)
102
70.2M
#define b0  bp
103
57.0M
#define b1  (bp + n)
104
105
34.1M
  s = an >> 1;
106
34.1M
  n = an - s;
107
34.1M
  t = bn - n;
108
109
34.1M
  ASSERT (an >= bn);
110
111
34.1M
  ASSERT (0 < s && s <= n && (n - s) == (an & 1));
112
34.1M
  ASSERT (0 < t && t <= s);
113
114
34.1M
  asm1 = pp;
115
34.1M
  bsm1 = pp + n;
116
117
34.1M
  vm1_neg = 0;
118
119
  /* Compute asm1.  */
120
34.1M
  if ((an & 1) == 0) /* s == n */
121
20.9M
    {
122
20.9M
      if (mpn_cmp (a0, a1, n) < 0)
123
8.65M
  {
124
8.65M
    mpn_sub_n (asm1, a1, a0, n);
125
8.65M
    vm1_neg = 1;
126
8.65M
  }
127
12.2M
      else
128
12.2M
  {
129
12.2M
    mpn_sub_n (asm1, a0, a1, n);
130
12.2M
  }
131
20.9M
    }
132
13.2M
  else /* n - s == 1 */
133
13.2M
    {
134
13.2M
      if (a0[s] == 0 && mpn_cmp (a0, a1, s) < 0)
135
224k
  {
136
224k
    mpn_sub_n (asm1, a1, a0, s);
137
224k
    asm1[s] = 0;
138
224k
    vm1_neg = 1;
139
224k
  }
140
12.9M
      else
141
12.9M
  {
142
12.9M
    asm1[s] = a0[s] - mpn_sub_n (asm1, a0, a1, s);
143
12.9M
  }
144
13.2M
    }
145
146
  /* Compute bsm1.  */
147
34.1M
  if (t == n)
148
20.9M
    {
149
20.9M
      if (mpn_cmp (b0, b1, n) < 0)
150
9.73M
  {
151
9.73M
    mpn_sub_n (bsm1, b1, b0, n);
152
9.73M
    vm1_neg ^= 1;
153
9.73M
  }
154
11.1M
      else
155
11.1M
  {
156
11.1M
    mpn_sub_n (bsm1, b0, b1, n);
157
11.1M
  }
158
20.9M
    }
159
13.2M
  else
160
13.2M
    {
161
13.2M
      if (mpn_zero_p (b0 + t, n - t) && mpn_cmp (b0, b1, t) < 0)
162
1.04M
  {
163
1.04M
    mpn_sub_n (bsm1, b1, b0, t);
164
1.04M
    MPN_ZERO (bsm1 + t, n - t);
165
1.04M
    vm1_neg ^= 1;
166
1.04M
  }
167
12.1M
      else
168
12.1M
  {
169
12.1M
    mpn_sub (bsm1, b0, n, b1, t);
170
12.1M
  }
171
13.2M
    }
172
173
68.3M
#define v0  pp        /* 2n */
174
68.3M
#define vinf  (pp + 2 * n)      /* s+t */
175
34.1M
#define vm1 scratch        /* 2n */
176
34.1M
#define scratch_out scratch + 2 * n
177
178
  /* vm1, 2n limbs */
179
34.1M
  TOOM22_MUL_N_REC (vm1, asm1, bsm1, n, scratch_out);
180
181
34.1M
  if (s > t)  TOOM22_MUL_REC (vinf, a1, s, b1, t, scratch_out);
182
34.1M
  else        TOOM22_MUL_N_REC (vinf, a1, b1, s, scratch_out);
183
184
  /* v0, 2n limbs */
185
34.1M
  TOOM22_MUL_N_REC (v0, ap, bp, n, scratch_out);
186
187
  /* H(v0) + L(vinf) */
188
34.1M
  cy = mpn_add_n (pp + 2 * n, v0 + n, vinf, n);
189
190
  /* L(v0) + (H(v0) + L(vinf)) */
191
34.1M
  cy2 = cy + mpn_add_n (pp + n, pp + 2 * n, v0, n);
192
193
  /* (H(v0) + L(vinf)) + H(vinf) */
194
34.1M
  cy += mpn_add (pp + 2 * n, pp + 2 * n, n, vinf + n, s + t - n);
195
196
34.1M
  if (vm1_neg)
197
9.98M
    cy += mpn_add_n (pp + n, pp + n, vm1, 2 * n);
198
24.1M
  else {
199
24.1M
    cy -= mpn_sub_n (pp + n, pp + n, vm1, 2 * n);
200
24.1M
    if (UNLIKELY (cy + 1 == 0)) { /* cy is negative */
201
      /* The total contribution of v0+vinf-vm1 can not be negative. */
202
#if WANT_ASSERT
203
      /* The borrow in cy stops the propagation of the carry cy2, */
204
      ASSERT (cy2 == 1);
205
      cy += mpn_add_1 (pp + 2 * n, pp + 2 * n, n, cy2);
206
      ASSERT (cy == 0);
207
#else
208
      /* we simply fill the area with zeros. */
209
3.01k
      MPN_FILL (pp + 2 * n, n, 0);
210
      /* ASSERT (s + t == n || mpn_zero_p (pp + 3 * n, s + t - n)); */
211
3.01k
#endif
212
3.01k
      return;
213
3.01k
    }
214
24.1M
  }
215
216
34.1M
  ASSERT (cy  <= 2);
217
34.1M
  ASSERT (cy2 <= 2);
218
219
34.1M
  MPN_INCR_U (pp + 2 * n, s + t, cy2);
220
  /* if s+t==n, cy is zero, but we should not access pp[3*n] at all. */
221
34.1M
  MPN_INCR_U (pp + 3 * n, s + t - n, cy);
222
34.1M
}