Coverage Report

Created: 2024-06-28 06:19

/src/gmp-6.2.1/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 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
177M
  (MUL_TOOM33_THRESHOLD >= 2 * MUL_TOOM22_THRESHOLD)
59
#endif
60
61
#define TOOM22_MUL_N_REC(p, a, b, n, ws)        \
62
88.9M
  do {                 \
63
88.9M
    if (! MAYBE_mul_toom22            \
64
88.9M
  || BELOW_THRESHOLD (n, MUL_TOOM22_THRESHOLD))     \
65
88.9M
      mpn_mul_basecase (p, a, n, b, n);         \
66
88.9M
    else                \
67
88.9M
      mpn_toom22_mul (p, a, n, b, n, ws);       \
68
88.9M
  } 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
3.61k
  do {                 \
78
3.61k
    if (! MAYBE_mul_toom22            \
79
3.61k
  || BELOW_THRESHOLD (bn, MUL_TOOM22_THRESHOLD))      \
80
3.61k
      mpn_mul_basecase (p, a, an, b, bn);       \
81
3.61k
    else if (4 * an < 5 * bn)           \
82
128
      mpn_toom22_mul (p, a, an, b, bn, ws);       \
83
128
    else                \
84
128
      mpn_toom32_mul (p, a, an, b, bn, ws);       \
85
3.61k
  } 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
29.6M
{
93
29.6M
  const int __gmpn_cpuvec_initialized = 1;
94
29.6M
  mp_size_t n, s, t;
95
29.6M
  int vm1_neg;
96
29.6M
  mp_limb_t cy, cy2;
97
29.6M
  mp_ptr asm1;
98
29.6M
  mp_ptr bsm1;
99
100
75.6M
#define a0  ap
101
43.7M
#define a1  (ap + n)
102
59.6M
#define b0  bp
103
43.7M
#define b1  (bp + n)
104
105
29.6M
  s = an >> 1;
106
29.6M
  n = an - s;
107
29.6M
  t = bn - n;
108
109
29.6M
  ASSERT (an >= bn);
110
111
29.6M
  ASSERT (0 < s && s <= n && s >= n - 1);
112
29.6M
  ASSERT (0 < t && t <= s);
113
114
29.6M
  asm1 = pp;
115
29.6M
  bsm1 = pp + n;
116
117
29.6M
  vm1_neg = 0;
118
119
  /* Compute asm1.  */
120
29.6M
  if (s == n)
121
13.7M
    {
122
13.7M
      if (mpn_cmp (a0, a1, n) < 0)
123
3.42M
  {
124
3.42M
    mpn_sub_n (asm1, a1, a0, n);
125
3.42M
    vm1_neg = 1;
126
3.42M
  }
127
10.2M
      else
128
10.2M
  {
129
10.2M
    mpn_sub_n (asm1, a0, a1, n);
130
10.2M
  }
131
13.7M
    }
132
15.9M
  else /* n - s == 1 */
133
15.9M
    {
134
15.9M
      if (a0[s] == 0 && mpn_cmp (a0, a1, s) < 0)
135
5.79k
  {
136
5.79k
    mpn_sub_n (asm1, a1, a0, s);
137
5.79k
    asm1[s] = 0;
138
5.79k
    vm1_neg = 1;
139
5.79k
  }
140
15.9M
      else
141
15.9M
  {
142
15.9M
    asm1[s] = a0[s] - mpn_sub_n (asm1, a0, a1, s);
143
15.9M
  }
144
15.9M
    }
145
146
  /* Compute bsm1.  */
147
29.6M
  if (t == n)
148
13.7M
    {
149
13.7M
      if (mpn_cmp (b0, b1, n) < 0)
150
3.72M
  {
151
3.72M
    mpn_sub_n (bsm1, b1, b0, n);
152
3.72M
    vm1_neg ^= 1;
153
3.72M
  }
154
9.99M
      else
155
9.99M
  {
156
9.99M
    mpn_sub_n (bsm1, b0, b1, n);
157
9.99M
  }
158
13.7M
    }
159
15.9M
  else
160
15.9M
    {
161
15.9M
      if (mpn_zero_p (b0 + t, n - t) && mpn_cmp (b0, b1, t) < 0)
162
297k
  {
163
297k
    mpn_sub_n (bsm1, b1, b0, t);
164
297k
    MPN_ZERO (bsm1 + t, n - t);
165
297k
    vm1_neg ^= 1;
166
297k
  }
167
15.6M
      else
168
15.6M
  {
169
15.6M
    mpn_sub (bsm1, b0, n, b1, t);
170
15.6M
  }
171
15.9M
    }
172
173
59.3M
#define v0  pp        /* 2n */
174
59.3M
#define vinf  (pp + 2 * n)      /* s+t */
175
29.6M
#define vm1 scratch        /* 2n */
176
29.6M
#define scratch_out scratch + 2 * n
177
178
  /* vm1, 2n limbs */
179
29.6M
  TOOM22_MUL_N_REC (vm1, asm1, bsm1, n, scratch_out);
180
181
29.6M
  if (s > t)  TOOM22_MUL_REC (vinf, a1, s, b1, t, scratch_out);
182
29.6M
  else        TOOM22_MUL_N_REC (vinf, a1, b1, s, scratch_out);
183
184
  /* v0, 2n limbs */
185
29.6M
  TOOM22_MUL_N_REC (v0, ap, bp, n, scratch_out);
186
187
  /* H(v0) + L(vinf) */
188
29.6M
  cy = mpn_add_n (pp + 2 * n, v0 + n, vinf, n);
189
190
  /* L(v0) + H(v0) */
191
29.6M
  cy2 = cy + mpn_add_n (pp + n, pp + 2 * n, v0, n);
192
193
  /* L(vinf) + H(vinf) */
194
29.6M
  cy += mpn_add (pp + 2 * n, pp + 2 * n, n, vinf + n, s + t - n);
195
196
29.6M
  if (vm1_neg)
197
3.97M
    cy += mpn_add_n (pp + n, pp + n, vm1, 2 * n);
198
25.6M
  else {
199
25.6M
    cy -= mpn_sub_n (pp + n, pp + n, vm1, 2 * n);
200
25.6M
    if (UNLIKELY (cy + 1 == 0)) { /* cy is negative */
201
      /* The total contribution of v0+vinf-vm1 can not be negative. */
202
8.16k
#if WANT_ASSERT
203
      /* The borrow in cy stops the propagation of the carry cy2, */
204
8.16k
      ASSERT (cy2 == 1);
205
8.16k
      cy += mpn_add_1 (pp + 2 * n, pp + 2 * n, n, cy2);
206
8.16k
      ASSERT (cy == 0);
207
#else
208
      /* we simply fill the area with zeros. */
209
      MPN_FILL (pp + 2 * n, n, 0);
210
#endif
211
8.16k
      return;
212
8.16k
    }
213
25.6M
  }
214
215
29.6M
  ASSERT (cy  <= 2);
216
29.6M
  ASSERT (cy2 <= 2);
217
218
29.6M
  MPN_INCR_U (pp + 2 * n, s + t, cy2);
219
  /* if s+t==n, cy is zero, but we should not access pp[3*n] at all. */
220
29.6M
  MPN_INCR_U (pp + 3 * n, s + t - n, cy);
221
29.6M
}