1 /* mpn_toom62_mul -- Multiply {ap,an} and {bp,bn} where an is nominally 3 times
2 as large as bn. Or more accurately, (5/2)bn < an < 6bn.
3
4 Contributed to the GNU project by Torbjorn Granlund and Marco Bodrato.
5
6 The idea of applying toom to unbalanced multiplication is due to Marco
7 Bodrato and Alberto Zanoni.
8
9 THE FUNCTION IN THIS FILE IS INTERNAL WITH A MUTABLE INTERFACE. IT IS ONLY
10 SAFE TO REACH IT THROUGH DOCUMENTED INTERFACES. IN FACT, IT IS ALMOST
11 GUARANTEED THAT IT WILL CHANGE OR DISAPPEAR IN A FUTURE GNU MP RELEASE.
12
13 Copyright 2006-2008, 2012 Free Software Foundation, Inc.
14
15 This file is part of the GNU MP Library.
16
17 The GNU MP Library is free software; you can redistribute it and/or modify
18 it under the terms of either:
19
20 * the GNU Lesser General Public License as published by the Free
21 Software Foundation; either version 3 of the License, or (at your
22 option) any later version.
23
24 or
25
26 * the GNU General Public License as published by the Free Software
27 Foundation; either version 2 of the License, or (at your option) any
28 later version.
29
30 or both in parallel, as here.
31
32 The GNU MP Library is distributed in the hope that it will be useful, but
33 WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
34 or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
35 for more details.
36
37 You should have received copies of the GNU General Public License and the
38 GNU Lesser General Public License along with the GNU MP Library. If not,
39 see https://www.gnu.org/licenses/. */
40
41
42 #include "gmp-impl.h"
43
44 /* Evaluate in:
45 0, +1, -1, +2, -2, 1/2, +inf
46
47 <-s-><--n--><--n--><--n--><--n--><--n-->
48 ___ ______ ______ ______ ______ ______
49 |a5_|___a4_|___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+ a4+ a5)*( b0+ b1) # A(1)*B(1) ah <= 5 bh <= 1
55 vm1 = ( a0- a1+ a2- a3+ a4- a5)*( b0- b1) # A(-1)*B(-1) |ah| <= 2 bh = 0
56 v2 = ( a0+ 2a1+4a2+8a3+16a4+32a5)*( b0+2b1) # A(2)*B(2) ah <= 62 bh <= 2
57 vm2 = ( a0- 2a1+4a2-8a3+16a4-32a5)*( b0-2b1) # A(-2)*B(-2) -41<=ah<=20 -1<=bh<=0
58 vh = (32a0+16a1+8a2+4a3+ 2a4+ a5)*(2b0+ b1) # A(1/2)*B(1/2) ah <= 62 bh <= 2
59 vinf= a5 * b1 # A(inf)*B(inf)
60 */
61
62 void
mpn_toom62_mul(mp_ptr pp,mp_srcptr ap,mp_size_t an,mp_srcptr bp,mp_size_t bn,mp_ptr scratch)63 mpn_toom62_mul (mp_ptr pp,
64 mp_srcptr ap, mp_size_t an,
65 mp_srcptr bp, mp_size_t bn,
66 mp_ptr scratch)
67 {
68 mp_size_t n, s, t;
69 mp_limb_t cy;
70 mp_ptr as1, asm1, as2, asm2, ash;
71 mp_ptr bs1, bsm1, bs2, bsm2, bsh;
72 mp_ptr gp;
73 enum toom7_flags aflags, bflags;
74 TMP_DECL;
75
76 #define a0 ap
77 #define a1 (ap + n)
78 #define a2 (ap + 2*n)
79 #define a3 (ap + 3*n)
80 #define a4 (ap + 4*n)
81 #define a5 (ap + 5*n)
82 #define b0 bp
83 #define b1 (bp + n)
84
85 n = 1 + (an >= 3 * bn ? (an - 1) / (size_t) 6 : (bn - 1) >> 1);
86
87 s = an - 5 * n;
88 t = bn - n;
89
90 ASSERT (0 < s && s <= n);
91 ASSERT (0 < t && t <= n);
92
93 TMP_MARK;
94
95 as1 = TMP_SALLOC_LIMBS (n + 1);
96 asm1 = TMP_SALLOC_LIMBS (n + 1);
97 as2 = TMP_SALLOC_LIMBS (n + 1);
98 asm2 = TMP_SALLOC_LIMBS (n + 1);
99 ash = TMP_SALLOC_LIMBS (n + 1);
100
101 bs1 = TMP_SALLOC_LIMBS (n + 1);
102 bsm1 = TMP_SALLOC_LIMBS (n);
103 bs2 = TMP_SALLOC_LIMBS (n + 1);
104 bsm2 = TMP_SALLOC_LIMBS (n + 1);
105 bsh = TMP_SALLOC_LIMBS (n + 1);
106
107 gp = pp;
108
109 /* Compute as1 and asm1. */
110 aflags = (enum toom7_flags) (toom7_w3_neg & mpn_toom_eval_pm1 (as1, asm1, 5, ap, n, s, gp));
111
112 /* Compute as2 and asm2. */
113 aflags = (enum toom7_flags) (aflags | (toom7_w1_neg & mpn_toom_eval_pm2 (as2, asm2, 5, ap, n, s, gp)));
114
115 /* Compute ash = 32 a0 + 16 a1 + 8 a2 + 4 a3 + 2 a4 + a5
116 = 2*(2*(2*(2*(2*a0 + a1) + a2) + a3) + a4) + a5 */
117
118 #if HAVE_NATIVE_mpn_addlsh1_n
119 cy = mpn_addlsh1_n (ash, a1, a0, n);
120 cy = 2*cy + mpn_addlsh1_n (ash, a2, ash, n);
121 cy = 2*cy + mpn_addlsh1_n (ash, a3, ash, n);
122 cy = 2*cy + mpn_addlsh1_n (ash, a4, ash, n);
123 if (s < n)
124 {
125 mp_limb_t cy2;
126 cy2 = mpn_addlsh1_n (ash, a5, ash, s);
127 ash[n] = 2*cy + mpn_lshift (ash + s, ash + s, n - s, 1);
128 MPN_INCR_U (ash + s, n+1-s, cy2);
129 }
130 else
131 ash[n] = 2*cy + mpn_addlsh1_n (ash, a5, ash, n);
132 #else
133 cy = mpn_lshift (ash, a0, n, 1);
134 cy += mpn_add_n (ash, ash, a1, n);
135 cy = 2*cy + mpn_lshift (ash, ash, n, 1);
136 cy += mpn_add_n (ash, ash, a2, n);
137 cy = 2*cy + mpn_lshift (ash, ash, n, 1);
138 cy += mpn_add_n (ash, ash, a3, n);
139 cy = 2*cy + mpn_lshift (ash, ash, n, 1);
140 cy += mpn_add_n (ash, ash, a4, n);
141 cy = 2*cy + mpn_lshift (ash, ash, n, 1);
142 ash[n] = cy + mpn_add (ash, ash, n, a5, s);
143 #endif
144
145 /* Compute bs1 and bsm1. */
146 if (t == n)
147 {
148 #if HAVE_NATIVE_mpn_add_n_sub_n
149 if (mpn_cmp (b0, b1, n) < 0)
150 {
151 cy = mpn_add_n_sub_n (bs1, bsm1, b1, b0, n);
152 bflags = toom7_w3_neg;
153 }
154 else
155 {
156 cy = mpn_add_n_sub_n (bs1, bsm1, b0, b1, n);
157 bflags = (enum toom7_flags) 0;
158 }
159 bs1[n] = cy >> 1;
160 #else
161 bs1[n] = mpn_add_n (bs1, b0, b1, n);
162 if (mpn_cmp (b0, b1, n) < 0)
163 {
164 mpn_sub_n (bsm1, b1, b0, n);
165 bflags = toom7_w3_neg;
166 }
167 else
168 {
169 mpn_sub_n (bsm1, b0, b1, n);
170 bflags = (enum toom7_flags) 0;
171 }
172 #endif
173 }
174 else
175 {
176 bs1[n] = mpn_add (bs1, b0, n, b1, t);
177 if (mpn_zero_p (b0 + t, n - t) && mpn_cmp (b0, b1, t) < 0)
178 {
179 mpn_sub_n (bsm1, b1, b0, t);
180 MPN_ZERO (bsm1 + t, n - t);
181 bflags = toom7_w3_neg;
182 }
183 else
184 {
185 mpn_sub (bsm1, b0, n, b1, t);
186 bflags = (enum toom7_flags) 0;
187 }
188 }
189
190 /* Compute bs2 and bsm2. Recycling bs1 and bsm1; bs2=bs1+b1, bsm2 =
191 bsm1 - b1 */
192 mpn_add (bs2, bs1, n + 1, b1, t);
193 if (bflags & toom7_w3_neg)
194 {
195 bsm2[n] = mpn_add (bsm2, bsm1, n, b1, t);
196 bflags = (enum toom7_flags) (bflags | toom7_w1_neg);
197 }
198 else
199 {
200 /* FIXME: Simplify this logic? */
201 if (t < n)
202 {
203 if (mpn_zero_p (bsm1 + t, n - t) && mpn_cmp (bsm1, b1, t) < 0)
204 {
205 ASSERT_NOCARRY (mpn_sub_n (bsm2, b1, bsm1, t));
206 MPN_ZERO (bsm2 + t, n + 1 - t);
207 bflags = (enum toom7_flags) (bflags | toom7_w1_neg);
208 }
209 else
210 {
211 ASSERT_NOCARRY (mpn_sub (bsm2, bsm1, n, b1, t));
212 bsm2[n] = 0;
213 }
214 }
215 else
216 {
217 if (mpn_cmp (bsm1, b1, n) < 0)
218 {
219 ASSERT_NOCARRY (mpn_sub_n (bsm2, b1, bsm1, n));
220 bflags = (enum toom7_flags) (bflags | toom7_w1_neg);
221 }
222 else
223 {
224 ASSERT_NOCARRY (mpn_sub_n (bsm2, bsm1, b1, n));
225 }
226 bsm2[n] = 0;
227 }
228 }
229
230 /* Compute bsh, recycling bs1. bsh=bs1+b0; */
231 bsh[n] = bs1[n] + mpn_add_n (bsh, bs1, b0, n);
232
233 ASSERT (as1[n] <= 5);
234 ASSERT (bs1[n] <= 1);
235 ASSERT (asm1[n] <= 2);
236 ASSERT (as2[n] <= 62);
237 ASSERT (bs2[n] <= 2);
238 ASSERT (asm2[n] <= 41);
239 ASSERT (bsm2[n] <= 1);
240 ASSERT (ash[n] <= 62);
241 ASSERT (bsh[n] <= 2);
242
243 #define v0 pp /* 2n */
244 #define v1 (pp + 2 * n) /* 2n+1 */
245 #define vinf (pp + 6 * n) /* s+t */
246 #define v2 scratch /* 2n+1 */
247 #define vm2 (scratch + 2 * n + 1) /* 2n+1 */
248 #define vh (scratch + 4 * n + 2) /* 2n+1 */
249 #define vm1 (scratch + 6 * n + 3) /* 2n+1 */
250 #define scratch_out (scratch + 8 * n + 4) /* 2n+1 */
251 /* Total scratch need: 10*n+5 */
252
253 /* Must be in allocation order, as they overwrite one limb beyond
254 * 2n+1. */
255 mpn_mul_n (v2, as2, bs2, n + 1); /* v2, 2n+1 limbs */
256 mpn_mul_n (vm2, asm2, bsm2, n + 1); /* vm2, 2n+1 limbs */
257 mpn_mul_n (vh, ash, bsh, n + 1); /* vh, 2n+1 limbs */
258
259 /* vm1, 2n+1 limbs */
260 mpn_mul_n (vm1, asm1, bsm1, n);
261 cy = 0;
262 if (asm1[n] == 1)
263 {
264 cy = mpn_add_n (vm1 + n, vm1 + n, bsm1, n);
265 }
266 else if (asm1[n] == 2)
267 {
268 #if HAVE_NATIVE_mpn_addlsh1_n
269 cy = mpn_addlsh1_n (vm1 + n, vm1 + n, bsm1, n);
270 #else
271 cy = mpn_addmul_1 (vm1 + n, bsm1, n, CNST_LIMB(2));
272 #endif
273 }
274 vm1[2 * n] = cy;
275
276 /* v1, 2n+1 limbs */
277 mpn_mul_n (v1, as1, bs1, n);
278 if (as1[n] == 1)
279 {
280 cy = bs1[n] + mpn_add_n (v1 + n, v1 + n, bs1, n);
281 }
282 else if (as1[n] == 2)
283 {
284 #if HAVE_NATIVE_mpn_addlsh1_n
285 cy = 2 * bs1[n] + mpn_addlsh1_n (v1 + n, v1 + n, bs1, n);
286 #else
287 cy = 2 * bs1[n] + mpn_addmul_1 (v1 + n, bs1, n, CNST_LIMB(2));
288 #endif
289 }
290 else if (as1[n] != 0)
291 {
292 cy = as1[n] * bs1[n] + mpn_addmul_1 (v1 + n, bs1, n, as1[n]);
293 }
294 else
295 cy = 0;
296 if (bs1[n] != 0)
297 cy += mpn_add_n (v1 + n, v1 + n, as1, n);
298 v1[2 * n] = cy;
299
300 mpn_mul_n (v0, a0, b0, n); /* v0, 2n limbs */
301
302 /* vinf, s+t limbs */
303 if (s > t) mpn_mul (vinf, a5, s, b1, t);
304 else mpn_mul (vinf, b1, t, a5, s);
305
306 mpn_toom_interpolate_7pts (pp, n, (enum toom7_flags) (aflags ^ bflags),
307 vm2, vm1, v2, vh, s + t, scratch_out);
308
309 TMP_FREE;
310 }
311