1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
|
/* spv.c - "small prime vector" functions for arithmetic on vectors of
residues modulo a single small prime
Copyright 2005, 2008 Dave Newman and Jason Papadopoulos.
The SP Library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or (at your
option) any later version.
The SP Library is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
License for more details.
You should have received a copy of the GNU Lesser General Public License
along with the SP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston,
MA 02110-1301, USA.
*/
#include <string.h> /* for memset */
#include "sp.h"
/* Routines for vectors of integers modulo r common small prime
*
* These are low-overhead routines that don't do memory allocation,
* other than for temporary variables. Unless otherwise specified, any
* of the input pointers can be equal. */
/* r = x */
void
spv_set (spv_t r, spv_t x, spv_size_t len)
{
#ifdef HAVE_MEMMOVE
/* memmove doesn't rely on the assertion below */
memmove (r, x, len * sizeof (sp_t));
#else
spv_size_t i;
ASSERT (r >= x + len || x >= r);
for (i = 0; i < len; i++)
r[i] = x[i];
#endif
}
/* r[0 ... len - 1] = x[len - 1 ... 0] */
void
spv_rev (spv_t r, spv_t x, spv_size_t len)
{
spv_size_t i;
ASSERT (r >= x + len || x >= r + len);
for (i = 0; i < len; i++)
r[i] = x[len - 1 - i];
}
/* r = [y, y, ... ] */
void
spv_set_sp (spv_t r, sp_t y, spv_size_t len)
{
spv_size_t i;
for (i = 0; i < len; i++)
r[i] = y;
}
void
spv_set_zero (spv_t r, spv_size_t len)
{
memset (r, 0, len * sizeof (sp_t));
}
int
spv_cmp (spv_t x, spv_t y, spv_size_t len)
{
spv_size_t i;
for (i = 0; i < len; i++)
if (x[i] != y[i])
return 1;
return 0;
}
/* r = x + y */
void
spv_add (spv_t r, spv_t x, spv_t y, spv_size_t len, sp_t m)
{
spv_size_t i;
ASSERT (r >= x + len || x >= r);
ASSERT (r >= y + len || y >= r);
for (i = 0; i < len; i++)
r[i] = sp_add (x[i], y[i], m);
}
/* r = [x[0] + y, x[1] + y, ... ] */
void
spv_add_sp (spv_t r, spv_t x, sp_t c, spv_size_t len, sp_t m)
{
spv_size_t i;
for (i = 0; i < len; i++)
r[i] = sp_add (x[i], c, m);
}
/* r = x - y */
void
spv_sub (spv_t r, spv_t x, spv_t y, spv_size_t len, sp_t m)
{
spv_size_t i;
ASSERT (r >= x + len || x >= r);
ASSERT (r >= y + len || y >= r);
for (i = 0; i < len; i++)
r[i] = sp_sub (x[i], y[i], m);
}
/* r = [x[0] - y, x[1] - y, ... ] */
void
spv_sub_sp (spv_t r, spv_t x, sp_t c, spv_size_t len, sp_t m)
{
spv_size_t i;
for (i = 0; i < len; i++)
r[i] = sp_sub (x[i], c, m);
}
/* r = [-x[0], -x[1], ... ] */
void
spv_neg (spv_t r, spv_t x, spv_size_t len, sp_t m)
{
spv_size_t i;
for (i = 0; i < len; i++)
r[i] = sp_sub (0, x[i], m);
}
/* Pointwise multiplication
* r = [x[0] * y[0], x[1] * y[1], ... ] */
void
spv_pwmul (spv_t r, spv_t x, spv_t y, spv_size_t len, sp_t m, sp_t d)
{
spv_size_t i = 0;
ASSERT (r >= x + len || x >= r);
ASSERT (r >= y + len || y >= r);
#if (defined(__GNUC__) || defined(__ICL)) && \
defined(__i386__) && defined(HAVE_SSE2)
asm volatile (
"movd %6, %%xmm6 \n\t"
"pshufd $0x44, %%xmm6, %%xmm5 \n\t"
"pshufd $0, %%xmm6, %%xmm6 \n\t"
"movd %7, %%xmm7 \n\t"
"pshufd $0, %%xmm7, %%xmm7 \n\t"
"0: \n\t"
"movdqa (%1,%4,4), %%xmm0 \n\t"
"movdqa (%2,%4,4), %%xmm2 \n\t"
"pshufd $0x31, %%xmm0, %%xmm1\n\t"
"pshufd $0x31, %%xmm2, %%xmm3\n\t"
"pmuludq %%xmm2, %%xmm0 \n\t"
"pmuludq %%xmm3, %%xmm1 \n\t"
"movdqa %%xmm0, %%xmm2 \n\t"
"movdqa %%xmm1, %%xmm3 \n\t"
"psrlq $" STRING((2*SP_NUMB_BITS - W_TYPE_SIZE)) ", %%xmm2 \n\t"
"pmuludq %%xmm7, %%xmm2 \n\t"
"psrlq $" STRING((2*SP_NUMB_BITS - W_TYPE_SIZE)) ", %%xmm3 \n\t"
"pmuludq %%xmm7, %%xmm3 \n\t"
#if SP_NUMB_BITS < W_TYPE_SIZE - 1
"psrlq $33, %%xmm2 \n\t"
"pmuludq %%xmm6, %%xmm2 \n\t"
"psrlq $33, %%xmm3 \n\t"
"pmuludq %%xmm6, %%xmm3 \n\t"
"psubq %%xmm2, %%xmm0 \n\t"
"psubq %%xmm3, %%xmm1 \n\t"
#else
"pshufd $0xf5, %%xmm2, %%xmm2 \n\t"
"pmuludq %%xmm6, %%xmm2 \n\t"
"pshufd $0xf5, %%xmm3, %%xmm3 \n\t"
"pmuludq %%xmm6, %%xmm3 \n\t"
"psubq %%xmm2, %%xmm0 \n\t"
"psubq %%xmm3, %%xmm1 \n\t"
"psubq %%xmm5, %%xmm0 \n\t"
"psubq %%xmm5, %%xmm1 \n\t"
"pshufd $0xf5, %%xmm0, %%xmm2 \n\t"
"pshufd $0xf5, %%xmm1, %%xmm3 \n\t"
"pand %%xmm5, %%xmm2 \n\t"
"pand %%xmm5, %%xmm3 \n\t"
"paddq %%xmm2, %%xmm0 \n\t"
"paddq %%xmm3, %%xmm1 \n\t"
#endif
"pshufd $0x8, %%xmm0, %%xmm0 \n\t"
"pshufd $0x8, %%xmm1, %%xmm1 \n\t"
"punpckldq %%xmm1, %%xmm0 \n\t"
"psubd %%xmm6, %%xmm0 \n\t"
"pxor %%xmm1, %%xmm1 \n\t"
"pcmpgtd %%xmm0, %%xmm1 \n\t"
"pand %%xmm6, %%xmm1 \n\t"
"paddd %%xmm1, %%xmm0 \n\t"
"movdqa %%xmm0, (%3,%4,4) \n\t"
"addl $4, %4 \n\t" /* INC */
"cmpl %5, %4 \n\t"
"jne 0b \n\t"
:"=r"(i)
:"r"(x), "r"(y), "r"(r), "0"(i),
"g"(len & (spv_size_t)(~3)), "g"(m), "g"(d)
:"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm5", "%xmm6", "%xmm7", "cc", "memory");
#elif defined( _MSC_VER ) && defined( SSE2)
__asm
{ push esi
push edi
mov edi, x
mov esi, y
mov edx, r
xor ecx, ecx
mov eax, len
and eax, ~3
movd xmm6, m
pshufd xmm5, xmm6, 0x44
pshufd xmm6, xmm6, 0
movd xmm7, d
pshufd xmm7, xmm7, 0
L0: movdqa xmm0, [edi+ecx*4]
movdqa xmm2, [esi+ecx*4]
pshufd xmm1, xmm0, 0x31
pshufd xmm3, xmm2, 0x31
pmuludq xmm0, xmm2
pmuludq xmm1, xmm3
movdqa xmm2, xmm0
movdqa xmm3, xmm1
psrlq xmm2, 2*SP_NUMB_BITS - W_TYPE_SIZE
pmuludq xmm2, xmm7
psrlq xmm3, 2*SP_NUMB_BITS - W_TYPE_SIZE
pmuludq xmm3, xmm7
#if SP_NUMB_BITS < W_TYPE_SIZE - 1
psrlq xmm2, 33
pmuludq xmm2, xmm6
psrlq xmm3, 33
pmuludq xmm3, xmm6
psubq xmm0, xmm2
psubq xmm1, xmm3
#else
pshufd xmm2, xmm2, 0xf5
pmuludq xmm2, xmm6
pshufd xmm3, xmm3, 0xf5
pmuludq xmm3, xmm6
psubq xmm0, xmm2
psubq xmm1, xmm3
psubq xmm0, xmm5
psubq xmm1, xmm5
pshufd xmm2, xmm0, 0xf5
pshufd xmm3, xmm1, 0xf5
pand xmm2, xmm5
pand xmm3, xmm5
paddq xmm0, xmm2
paddq xmm1, xmm3
#endif
pshufd xmm0, xmm0, 0x8
pshufd xmm1, xmm1, 0x8
punpckldq xmm0, xmm1
psubd xmm0, xmm6
pxor xmm1, xmm1
pcmpgtd xmm1, xmm0
pand xmm1, xmm6
paddd xmm0, xmm1
movdqa [edx+ecx*4], xmm0
add ecx, 4
cmp eax, ecx
jne L0
mov i, ecx
pop edi
pop esi
}
#endif
for (; i < len; i++)
r[i] = sp_mul (x[i], y[i], m, d);
}
/* Pointwise multiplication, second input is read in reverse
* r = [x[0] * y[len - 1], x[1] * y[len - 2], ... x[len - 1] * y[0]] */
void
spv_pwmul_rev (spv_t r, spv_t x, spv_t y, spv_size_t len, sp_t m, sp_t d)
{
spv_size_t i;
ASSERT (r >= x + len || x >= r);
ASSERT (r >= y + len || y >= r);
for (i = 0; i < len; i++)
r[i] = sp_mul (x[i], y[len - 1 - i], m, d);
}
/* dst = src * y */
void
spv_mul_sp (spv_t r, spv_t x, sp_t c, spv_size_t len, sp_t m, sp_t d)
{
spv_size_t i = 0;
ASSERT (r >= x + len || x >= r);
#if (defined(__GNUC__) || defined(__ICL)) && \
defined(__i386__) && defined(HAVE_SSE2)
asm volatile (
"movd %2, %%xmm4 \n\t"
"pshufd $0, %%xmm4, %%xmm4 \n\t"
"movd %6, %%xmm6 \n\t"
"pshufd $0x44, %%xmm6, %%xmm5 \n\t"
"pshufd $0, %%xmm6, %%xmm6 \n\t"
"movd %7, %%xmm7 \n\t"
"pshufd $0, %%xmm7, %%xmm7 \n\t"
"0: \n\t"
"movdqa (%1,%4,4), %%xmm0 \n\t"
"pshufd $0x31, %%xmm0, %%xmm1\n\t"
"pshufd $0x31, %%xmm4, %%xmm3\n\t"
"pmuludq %%xmm4, %%xmm0 \n\t"
"pmuludq %%xmm3, %%xmm1 \n\t"
"movdqa %%xmm0, %%xmm2 \n\t"
"movdqa %%xmm1, %%xmm3 \n\t"
"psrlq $" STRING((2*SP_NUMB_BITS - W_TYPE_SIZE)) ", %%xmm2 \n\t"
"pmuludq %%xmm7, %%xmm2 \n\t"
"psrlq $" STRING((2*SP_NUMB_BITS - W_TYPE_SIZE)) ", %%xmm3 \n\t"
"pmuludq %%xmm7, %%xmm3 \n\t"
#if SP_NUMB_BITS < W_TYPE_SIZE - 1
"psrlq $33, %%xmm2 \n\t"
"pmuludq %%xmm6, %%xmm2 \n\t"
"psrlq $33, %%xmm3 \n\t"
"pmuludq %%xmm6, %%xmm3 \n\t"
"psubq %%xmm2, %%xmm0 \n\t"
"psubq %%xmm3, %%xmm1 \n\t"
#else
"pshufd $0xf5, %%xmm2, %%xmm2 \n\t"
"pmuludq %%xmm6, %%xmm2 \n\t"
"pshufd $0xf5, %%xmm3, %%xmm3 \n\t"
"pmuludq %%xmm6, %%xmm3 \n\t"
"psubq %%xmm2, %%xmm0 \n\t"
"psubq %%xmm3, %%xmm1 \n\t"
"psubq %%xmm5, %%xmm0 \n\t"
"psubq %%xmm5, %%xmm1 \n\t"
"pshufd $0xf5, %%xmm0, %%xmm2 \n\t"
"pshufd $0xf5, %%xmm1, %%xmm3 \n\t"
"pand %%xmm5, %%xmm2 \n\t"
"pand %%xmm5, %%xmm3 \n\t"
"paddq %%xmm2, %%xmm0 \n\t"
"paddq %%xmm3, %%xmm1 \n\t"
#endif
"pshufd $0x8, %%xmm0, %%xmm0 \n\t"
"pshufd $0x8, %%xmm1, %%xmm1 \n\t"
"punpckldq %%xmm1, %%xmm0 \n\t"
"psubd %%xmm6, %%xmm0 \n\t"
"pxor %%xmm1, %%xmm1 \n\t"
"pcmpgtd %%xmm0, %%xmm1 \n\t"
"pand %%xmm6, %%xmm1 \n\t"
"paddd %%xmm1, %%xmm0 \n\t"
"movdqa %%xmm0, (%3,%4,4) \n\t"
"addl $4, %4 \n\t" /* INC */
"cmpl %5, %4 \n\t"
"jne 0b \n\t"
:"=r"(i)
:"r"(x), "g"(c), "r"(r), "0"(i),
"g"(len & (spv_size_t)(~3)), "g"(m), "g"(d)
:"%xmm0", "%xmm1", "%xmm2", "%xmm3",
"%xmm4", "%xmm5", "%xmm6", "%xmm7", "cc", "memory");
#elif defined( _MSC_VER ) && defined( SSE2)
__asm
{ push esi
push edi
xor ecx, ecx
mov edi, x
mov esi, c
mov edx, r
mov eax, len
and eax, ~3
movd xmm4, esi
pshufd xmm4, xmm4, 0
movd xmm6, m
pshufd xmm5, xmm6, 0x44
pshufd xmm6, xmm6, 0
movd xmm7, d
pshufd xmm7, xmm7, 0
L0: movdqa xmm0, [edi+ecx*4]
pshufd xmm1, xmm0, 0x31
pshufd xmm3, xmm4, 0x31
pmuludq xmm0, xmm4
pmuludq xmm1, xmm3
movdqa xmm2, xmm0
movdqa xmm3, xmm1
psrlq xmm2, 2*SP_NUMB_BITS - W_TYPE_SIZE
pmuludq xmm2, xmm7
psrlq xmm3, 2*SP_NUMB_BITS - W_TYPE_SIZE
pmuludq xmm3, xmm7
#if SP_NUMB_BITS < W_TYPE_SIZE - 1
psrlq xmm2, 33
pmuludq xmm2, xmm6
psrlq xmm3, 33
pmuludq xmm3, xmm6
psubq xmm0, xmm2
psubq xmm1, xmm3
#else
pshufd xmm2, xmm2, 0xf5
pmuludq xmm2, xmm6
pshufd xmm3, xmm3, 0xf5
pmuludq xmm3, xmm6
psubq xmm0, xmm2
psubq xmm1, xmm3
psubq xmm0, xmm5
psubq xmm1, xmm5
pshufd xmm2, xmm0, 0xf5
pshufd xmm3, xmm1, 0xf5
pand xmm2, xmm5
pand xmm3, xmm5
paddq xmm0, xmm2
paddq xmm1, xmm3
#endif
pshufd xmm0, xmm0, 0x8
pshufd xmm1, xmm1, 0x8
punpckldq xmm0, xmm1
psubd xmm0, xmm6
pxor xmm1, xmm1
pcmpgtd xmm1, xmm0
pand xmm1, xmm6
paddd xmm0, xmm1
movdqa [edx+ecx*4], xmm0
add ecx, 4
cmp eax, ecx
jne L0
mov i, ecx
pop edi
pop esi
}
#endif
for (; i < len; i++)
r[i] = sp_mul (x[i], c, m, d);
}
void
spv_random (spv_t x, spv_size_t len, sp_t m)
{
spv_size_t i;
mpn_random (x, len);
for (i = 0; i < len; i++)
while (x[i] >= m)
x[i] -= m;
}
|
