1 /* mpn_toom42_mul -- Multiply {ap,an} and {bp,bn} where an is nominally twice
2 as large as bn. Or more accurately, (3/2)bn < an < 4bn.
3
4 Contributed to the GNU project by Torbjorn Granlund.
5 Additional improvements by Marco Bodrato.
6
7 The idea of applying toom to unbalanced multiplication is due to Marco
8 Bodrato and Alberto Zanoni.
9
10 THE FUNCTION IN THIS FILE IS INTERNAL WITH A MUTABLE INTERFACE. IT IS ONLY
11 SAFE TO REACH IT THROUGH DOCUMENTED INTERFACES. IN FACT, IT IS ALMOST
12 GUARANTEED THAT IT WILL CHANGE OR DISAPPEAR IN A FUTURE GNU MP RELEASE.
13
14 Copyright 2006-2008, 2012, 2014 Free Software Foundation, Inc.
15
16 This file is part of the GNU MP Library.
17
18 The GNU MP Library is free software; you can redistribute it and/or modify
19 it under the terms of either:
20
21 * the GNU Lesser General Public License as published by the Free
22 Software Foundation; either version 3 of the License, or (at your
23 option) any later version.
24
25 or
26
27 * the GNU General Public License as published by the Free Software
28 Foundation; either version 2 of the License, or (at your option) any
29 later version.
30
31 or both in parallel, as here.
32
33 The GNU MP Library is distributed in the hope that it will be useful, but
34 WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
35 or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
36 for more details.
37
38 You should have received copies of the GNU General Public License and the
39 GNU Lesser General Public License along with the GNU MP Library. If not,
40 see https://www.gnu.org/licenses/. */
41
42
43 #include "gmp-impl.h"
44
45 /* Evaluate in: -1, 0, +1, +2, +inf
46
47 <-s-><--n--><--n--><--n-->
48 ___ ______ ______ ______
49 |a3_|___a2_|___a1_|___a0_|
50 |_b1_|___b0_|
51 <-t--><--n-->
52
53 v0 = a0 * b0 # A(0)*B(0)
54 v1 = (a0+ a1+ a2+ a3)*(b0+ b1) # A(1)*B(1) ah <= 3 bh <= 1
55 vm1 = (a0- a1+ a2- a3)*(b0- b1) # A(-1)*B(-1) |ah| <= 1 bh = 0
56 v2 = (a0+2a1+4a2+8a3)*(b0+2b1) # A(2)*B(2) ah <= 14 bh <= 2
57 vinf= a3 * b1 # A(inf)*B(inf)
58 */
59
60 #define TOOM42_MUL_N_REC(p, a, b, n, ws) \
61 do { \
62 mpn_mul_n (p, a, b, n); \
63 } while (0)
64
65 void
mpn_toom42_mul(mp_ptr pp,mp_srcptr ap,mp_size_t an,mp_srcptr bp,mp_size_t bn,mp_ptr scratch)66 mpn_toom42_mul (mp_ptr pp,
67 mp_srcptr ap, mp_size_t an,
68 mp_srcptr bp, mp_size_t bn,
69 mp_ptr scratch)
70 {
71 mp_size_t n, s, t;
72 int vm1_neg;
73 mp_limb_t cy, vinf0;
74 mp_ptr a0_a2;
75 mp_ptr as1, asm1, as2;
76 mp_ptr bs1, bsm1, bs2;
77 mp_ptr tmp;
78 TMP_DECL;
79
80 #define a0 ap
81 #define a1 (ap + n)
82 #define a2 (ap + 2*n)
83 #define a3 (ap + 3*n)
84 #define b0 bp
85 #define b1 (bp + n)
86
87 n = an >= 2 * bn ? (an + 3) >> 2 : (bn + 1) >> 1;
88
89 s = an - 3 * n;
90 t = bn - n;
91
92 ASSERT (0 < s && s <= n);
93 ASSERT (0 < t && t <= n);
94
95 TMP_MARK;
96
97 tmp = TMP_ALLOC_LIMBS (6 * n + 5);
98 as1 = tmp; tmp += n + 1;
99 asm1 = tmp; tmp += n + 1;
100 as2 = tmp; tmp += n + 1;
101 bs1 = tmp; tmp += n + 1;
102 bsm1 = tmp; tmp += n;
103 bs2 = tmp; tmp += n + 1;
104
105 a0_a2 = pp;
106
107 /* Compute as1 and asm1. */
108 vm1_neg = mpn_toom_eval_dgr3_pm1 (as1, asm1, ap, n, s, a0_a2) & 1;
109
110 /* Compute as2. */
111 #if HAVE_NATIVE_mpn_addlsh1_n
112 cy = mpn_addlsh1_n (as2, a2, a3, s);
113 if (s != n)
114 cy = mpn_add_1 (as2 + s, a2 + s, n - s, cy);
115 cy = 2 * cy + mpn_addlsh1_n (as2, a1, as2, n);
116 cy = 2 * cy + mpn_addlsh1_n (as2, a0, as2, n);
117 #else
118 cy = mpn_lshift (as2, a3, s, 1);
119 cy += mpn_add_n (as2, a2, as2, s);
120 if (s != n)
121 cy = mpn_add_1 (as2 + s, a2 + s, n - s, cy);
122 cy = 2 * cy + mpn_lshift (as2, as2, n, 1);
123 cy += mpn_add_n (as2, a1, as2, n);
124 cy = 2 * cy + mpn_lshift (as2, as2, n, 1);
125 cy += mpn_add_n (as2, a0, as2, n);
126 #endif
127 as2[n] = cy;
128
129 /* Compute bs1 and bsm1. */
130 if (t == n)
131 {
132 #if HAVE_NATIVE_mpn_add_n_sub_n
133 if (mpn_cmp (b0, b1, n) < 0)
134 {
135 cy = mpn_add_n_sub_n (bs1, bsm1, b1, b0, n);
136 vm1_neg ^= 1;
137 }
138 else
139 {
140 cy = mpn_add_n_sub_n (bs1, bsm1, b0, b1, n);
141 }
142 bs1[n] = cy >> 1;
143 #else
144 bs1[n] = mpn_add_n (bs1, b0, b1, n);
145
146 if (mpn_cmp (b0, b1, n) < 0)
147 {
148 mpn_sub_n (bsm1, b1, b0, n);
149 vm1_neg ^= 1;
150 }
151 else
152 {
153 mpn_sub_n (bsm1, b0, b1, n);
154 }
155 #endif
156 }
157 else
158 {
159 bs1[n] = mpn_add (bs1, b0, n, b1, t);
160
161 if (mpn_zero_p (b0 + t, n - t) && mpn_cmp (b0, b1, t) < 0)
162 {
163 mpn_sub_n (bsm1, b1, b0, t);
164 MPN_ZERO (bsm1 + t, n - t);
165 vm1_neg ^= 1;
166 }
167 else
168 {
169 mpn_sub (bsm1, b0, n, b1, t);
170 }
171 }
172
173 /* Compute bs2, recycling bs1. bs2=bs1+b1 */
174 mpn_add (bs2, bs1, n + 1, b1, t);
175
176 ASSERT (as1[n] <= 3);
177 ASSERT (bs1[n] <= 1);
178 ASSERT (asm1[n] <= 1);
179 /*ASSERT (bsm1[n] == 0);*/
180 ASSERT (as2[n] <= 14);
181 ASSERT (bs2[n] <= 2);
182
183 #define v0 pp /* 2n */
184 #define v1 (pp + 2 * n) /* 2n+1 */
185 #define vinf (pp + 4 * n) /* s+t */
186 #define vm1 scratch /* 2n+1 */
187 #define v2 (scratch + 2 * n + 1) /* 2n+2 */
188 #define scratch_out scratch + 4 * n + 4 /* Currently unused. */
189
190 /* vm1, 2n+1 limbs */
191 TOOM42_MUL_N_REC (vm1, asm1, bsm1, n, scratch_out);
192 cy = 0;
193 if (asm1[n] != 0)
194 cy = mpn_add_n (vm1 + n, vm1 + n, bsm1, n);
195 vm1[2 * n] = cy;
196
197 TOOM42_MUL_N_REC (v2, as2, bs2, n + 1, scratch_out); /* v2, 2n+1 limbs */
198
199 /* vinf, s+t limbs */
200 if (s > t) mpn_mul (vinf, a3, s, b1, t);
201 else mpn_mul (vinf, b1, t, a3, s);
202
203 vinf0 = vinf[0]; /* v1 overlaps with this */
204
205 /* v1, 2n+1 limbs */
206 TOOM42_MUL_N_REC (v1, as1, bs1, n, scratch_out);
207 if (as1[n] == 1)
208 {
209 cy = bs1[n] + mpn_add_n (v1 + n, v1 + n, bs1, n);
210 }
211 else if (as1[n] == 2)
212 {
213 #if HAVE_NATIVE_mpn_addlsh1_n
214 cy = 2 * bs1[n] + mpn_addlsh1_n (v1 + n, v1 + n, bs1, n);
215 #else
216 cy = 2 * bs1[n] + mpn_addmul_1 (v1 + n, bs1, n, CNST_LIMB(2));
217 #endif
218 }
219 else if (as1[n] == 3)
220 {
221 cy = 3 * bs1[n] + mpn_addmul_1 (v1 + n, bs1, n, CNST_LIMB(3));
222 }
223 else
224 cy = 0;
225 if (bs1[n] != 0)
226 cy += mpn_add_n (v1 + n, v1 + n, as1, n);
227 v1[2 * n] = cy;
228
229 TOOM42_MUL_N_REC (v0, ap, bp, n, scratch_out); /* v0, 2n limbs */
230
231 mpn_toom_interpolate_5pts (pp, v2, vm1, n, s + t, vm1_neg, vinf0);
232
233 TMP_FREE;
234 }
235