1 |
/** |
2 |
* @file dSFMT.c |
3 |
* @brief double precision SIMD-oriented Fast Mersenne Twister (dSFMT) |
4 |
* based on IEEE 754 format. |
5 |
* |
6 |
* @author Mutsuo Saito (Hiroshima University) |
7 |
* @author Makoto Matsumoto (Hiroshima University) |
8 |
* |
9 |
* Copyright (C) 2007,2008 Mutsuo Saito, Makoto Matsumoto and Hiroshima |
10 |
* University. All rights reserved. |
11 |
* |
12 |
* The new BSD License is applied to this software, see LICENSE.txt |
13 |
*/ |
14 |
#include <stdio.h> |
15 |
#include <string.h> |
16 |
#include <stdlib.h> |
17 |
#include "dSFMT-params.h" |
18 |
|
19 |
/** dsfmt internal state vector */ |
20 |
dsfmt_t dsfmt_global_data; |
21 |
/** dsfmt mexp for check */ |
22 |
static const int dsfmt_mexp = DSFMT_MEXP; |
23 |
|
24 |
/*---------------- |
25 |
STATIC FUNCTIONS |
26 |
----------------*/ |
27 |
inline static uint32_t ini_func1(uint32_t x); |
28 |
inline static uint32_t ini_func2(uint32_t x); |
29 |
inline static void gen_rand_array_c1o2(dsfmt_t *dsfmt, w128_t *array, |
30 |
int size); |
31 |
inline static void gen_rand_array_c0o1(dsfmt_t *dsfmt, w128_t *array, |
32 |
int size); |
33 |
inline static void gen_rand_array_o0c1(dsfmt_t *dsfmt, w128_t *array, |
34 |
int size); |
35 |
inline static void gen_rand_array_o0o1(dsfmt_t *dsfmt, w128_t *array, |
36 |
int size); |
37 |
inline static int idxof(int i); |
38 |
static void initial_mask(dsfmt_t *dsfmt); |
39 |
static void period_certification(dsfmt_t *dsfmt); |
40 |
|
41 |
#if defined(HAVE_SSE2) |
42 |
# include <emmintrin.h> |
43 |
/** mask data for sse2 */ |
44 |
static __m128i sse2_param_mask; |
45 |
/** 1 in 64bit for sse2 */ |
46 |
static __m128i sse2_int_one; |
47 |
/** 2.0 double for sse2 */ |
48 |
static __m128d sse2_double_two; |
49 |
/** -1.0 double for sse2 */ |
50 |
static __m128d sse2_double_m_one; |
51 |
|
52 |
static void setup_const(void); |
53 |
#endif |
54 |
|
55 |
/** |
56 |
* This function simulate a 32-bit array index overlapped to 64-bit |
57 |
* array of LITTLE ENDIAN in BIG ENDIAN machine. |
58 |
*/ |
59 |
#if defined(DSFMT_BIG_ENDIAN) |
60 |
inline static int idxof(int i) { |
61 |
return i ^ 1; |
62 |
} |
63 |
#else |
64 |
inline static int idxof(int i) { |
65 |
return i; |
66 |
} |
67 |
#endif |
68 |
|
69 |
/** |
70 |
* This function represents the recursion formula. |
71 |
* @param r output |
72 |
* @param a a 128-bit part of the internal state array |
73 |
* @param b a 128-bit part of the internal state array |
74 |
* @param lung a 128-bit part of the internal state array |
75 |
*/ |
76 |
#if defined(HAVE_ALTIVEC) |
77 |
inline static void do_recursion(w128_t *r, w128_t *a, w128_t * b, |
78 |
w128_t *lung) { |
79 |
const vector unsigned char sl1 = ALTI_SL1; |
80 |
const vector unsigned char sl1_perm = ALTI_SL1_PERM; |
81 |
const vector unsigned int sl1_msk = ALTI_SL1_MSK; |
82 |
const vector unsigned char sr1 = ALTI_SR; |
83 |
const vector unsigned char sr1_perm = ALTI_SR_PERM; |
84 |
const vector unsigned int sr1_msk = ALTI_SR_MSK; |
85 |
const vector unsigned char perm = ALTI_PERM; |
86 |
const vector unsigned int msk1 = ALTI_MSK; |
87 |
vector unsigned int w, x, y, z; |
88 |
|
89 |
z = a->s; |
90 |
w = lung->s; |
91 |
x = vec_perm(w, (vector unsigned int)perm, perm); |
92 |
y = vec_perm(z, sl1_perm, sl1_perm); |
93 |
y = vec_sll(y, sl1); |
94 |
y = vec_and(y, sl1_msk); |
95 |
w = vec_xor(x, b->s); |
96 |
w = vec_xor(w, y); |
97 |
x = vec_perm(w, (vector unsigned int)sr1_perm, sr1_perm); |
98 |
x = vec_srl(x, sr1); |
99 |
x = vec_and(x, sr1_msk); |
100 |
y = vec_and(w, msk1); |
101 |
z = vec_xor(z, y); |
102 |
r->s = vec_xor(z, x); |
103 |
lung->s = w; |
104 |
} |
105 |
#elif defined(HAVE_SSE2) |
106 |
/** |
107 |
* This function setup some constant variables for SSE2. |
108 |
*/ |
109 |
static void setup_const(void) { |
110 |
static int first = 1; |
111 |
if (!first) { |
112 |
return; |
113 |
} |
114 |
sse2_param_mask = _mm_set_epi32(DSFMT_MSK32_3, DSFMT_MSK32_4, |
115 |
DSFMT_MSK32_1, DSFMT_MSK32_2); |
116 |
sse2_int_one = _mm_set_epi32(0, 1, 0, 1); |
117 |
sse2_double_two = _mm_set_pd(2.0, 2.0); |
118 |
sse2_double_m_one = _mm_set_pd(-1.0, -1.0); |
119 |
first = 0; |
120 |
} |
121 |
|
122 |
/** |
123 |
* This function represents the recursion formula. |
124 |
* @param r output 128-bit |
125 |
* @param a a 128-bit part of the internal state array |
126 |
* @param b a 128-bit part of the internal state array |
127 |
* @param d a 128-bit part of the internal state array (I/O) |
128 |
*/ |
129 |
inline static void do_recursion(w128_t *r, w128_t *a, w128_t *b, w128_t *u) { |
130 |
__m128i v, w, x, y, z; |
131 |
|
132 |
x = a->si; |
133 |
z = _mm_slli_epi64(x, DSFMT_SL1); |
134 |
y = _mm_shuffle_epi32(u->si, SSE2_SHUFF); |
135 |
z = _mm_xor_si128(z, b->si); |
136 |
y = _mm_xor_si128(y, z); |
137 |
|
138 |
v = _mm_srli_epi64(y, DSFMT_SR); |
139 |
w = _mm_and_si128(y, sse2_param_mask); |
140 |
v = _mm_xor_si128(v, x); |
141 |
v = _mm_xor_si128(v, w); |
142 |
r->si = v; |
143 |
u->si = y; |
144 |
} |
145 |
#else /* standard C */ |
146 |
/** |
147 |
* This function represents the recursion formula. |
148 |
* @param r output 128-bit |
149 |
* @param a a 128-bit part of the internal state array |
150 |
* @param b a 128-bit part of the internal state array |
151 |
* @param lung a 128-bit part of the internal state array (I/O) |
152 |
*/ |
153 |
inline static void do_recursion(w128_t *r, w128_t *a, w128_t * b, |
154 |
w128_t *lung) { |
155 |
uint64_t t0, t1, L0, L1; |
156 |
|
157 |
t0 = a->u[0]; |
158 |
t1 = a->u[1]; |
159 |
L0 = lung->u[0]; |
160 |
L1 = lung->u[1]; |
161 |
lung->u[0] = (t0 << DSFMT_SL1) ^ (L1 >> 32) ^ (L1 << 32) ^ b->u[0]; |
162 |
lung->u[1] = (t1 << DSFMT_SL1) ^ (L0 >> 32) ^ (L0 << 32) ^ b->u[1]; |
163 |
r->u[0] = (lung->u[0] >> DSFMT_SR) ^ (lung->u[0] & DSFMT_MSK1) ^ t0; |
164 |
r->u[1] = (lung->u[1] >> DSFMT_SR) ^ (lung->u[1] & DSFMT_MSK2) ^ t1; |
165 |
} |
166 |
#endif |
167 |
|
168 |
#if defined(HAVE_SSE2) |
169 |
/** |
170 |
* This function converts the double precision floating point numbers which |
171 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
172 |
* in the range [0, 1). |
173 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
174 |
*/ |
175 |
inline static void convert_c0o1(w128_t *w) { |
176 |
w->sd = _mm_add_pd(w->sd, sse2_double_m_one); |
177 |
} |
178 |
|
179 |
/** |
180 |
* This function converts the double precision floating point numbers which |
181 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
182 |
* in the range (0, 1]. |
183 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
184 |
*/ |
185 |
inline static void convert_o0c1(w128_t *w) { |
186 |
w->sd = _mm_sub_pd(sse2_double_two, w->sd); |
187 |
} |
188 |
|
189 |
/** |
190 |
* This function converts the double precision floating point numbers which |
191 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
192 |
* in the range (0, 1). |
193 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
194 |
*/ |
195 |
inline static void convert_o0o1(w128_t *w) { |
196 |
w->si = _mm_or_si128(w->si, sse2_int_one); |
197 |
w->sd = _mm_add_pd(w->sd, sse2_double_m_one); |
198 |
} |
199 |
#else /* standard C and altivec */ |
200 |
/** |
201 |
* This function converts the double precision floating point numbers which |
202 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
203 |
* in the range [0, 1). |
204 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
205 |
*/ |
206 |
inline static void convert_c0o1(w128_t *w) { |
207 |
w->d[0] -= 1.0; |
208 |
w->d[1] -= 1.0; |
209 |
} |
210 |
|
211 |
/** |
212 |
* This function converts the double precision floating point numbers which |
213 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
214 |
* in the range (0, 1]. |
215 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
216 |
*/ |
217 |
inline static void convert_o0c1(w128_t *w) { |
218 |
w->d[0] = 2.0 - w->d[0]; |
219 |
w->d[1] = 2.0 - w->d[1]; |
220 |
} |
221 |
|
222 |
/** |
223 |
* This function converts the double precision floating point numbers which |
224 |
* distribute uniformly in the range [1, 2) to those which distribute uniformly |
225 |
* in the range (0, 1). |
226 |
* @param w 128bit stracture of double precision floating point numbers (I/O) |
227 |
*/ |
228 |
inline static void convert_o0o1(w128_t *w) { |
229 |
w->u[0] |= 1; |
230 |
w->u[1] |= 1; |
231 |
w->d[0] -= 1.0; |
232 |
w->d[1] -= 1.0; |
233 |
} |
234 |
#endif |
235 |
|
236 |
/** |
237 |
* This function fills the user-specified array with double precision |
238 |
* floating point pseudorandom numbers of the IEEE 754 format. |
239 |
* @param dsfmt dsfmt state vector. |
240 |
* @param array an 128-bit array to be filled by pseudorandom numbers. |
241 |
* @param size number of 128-bit pseudorandom numbers to be generated. |
242 |
*/ |
243 |
inline static void gen_rand_array_c1o2(dsfmt_t *dsfmt, w128_t *array, |
244 |
int size) { |
245 |
int i, j; |
246 |
w128_t lung; |
247 |
|
248 |
lung = dsfmt->status[DSFMT_N]; |
249 |
do_recursion(&array[0], &dsfmt->status[0], &dsfmt->status[DSFMT_POS1], |
250 |
&lung); |
251 |
for (i = 1; i < DSFMT_N - DSFMT_POS1; i++) { |
252 |
do_recursion(&array[i], &dsfmt->status[i], |
253 |
&dsfmt->status[i + DSFMT_POS1], &lung); |
254 |
} |
255 |
for (; i < DSFMT_N; i++) { |
256 |
do_recursion(&array[i], &dsfmt->status[i], |
257 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
258 |
} |
259 |
for (; i < size - DSFMT_N; i++) { |
260 |
do_recursion(&array[i], &array[i - DSFMT_N], |
261 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
262 |
} |
263 |
for (j = 0; j < 2 * DSFMT_N - size; j++) { |
264 |
dsfmt->status[j] = array[j + size - DSFMT_N]; |
265 |
} |
266 |
for (; i < size; i++, j++) { |
267 |
do_recursion(&array[i], &array[i - DSFMT_N], |
268 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
269 |
dsfmt->status[j] = array[i]; |
270 |
} |
271 |
dsfmt->status[DSFMT_N] = lung; |
272 |
} |
273 |
|
274 |
/** |
275 |
* This function fills the user-specified array with double precision |
276 |
* floating point pseudorandom numbers of the IEEE 754 format. |
277 |
* @param dsfmt dsfmt state vector. |
278 |
* @param array an 128-bit array to be filled by pseudorandom numbers. |
279 |
* @param size number of 128-bit pseudorandom numbers to be generated. |
280 |
*/ |
281 |
inline static void gen_rand_array_c0o1(dsfmt_t *dsfmt, w128_t *array, |
282 |
int size) { |
283 |
int i, j; |
284 |
w128_t lung; |
285 |
|
286 |
lung = dsfmt->status[DSFMT_N]; |
287 |
do_recursion(&array[0], &dsfmt->status[0], &dsfmt->status[DSFMT_POS1], |
288 |
&lung); |
289 |
for (i = 1; i < DSFMT_N - DSFMT_POS1; i++) { |
290 |
do_recursion(&array[i], &dsfmt->status[i], |
291 |
&dsfmt->status[i + DSFMT_POS1], &lung); |
292 |
} |
293 |
for (; i < DSFMT_N; i++) { |
294 |
do_recursion(&array[i], &dsfmt->status[i], |
295 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
296 |
} |
297 |
for (; i < size - DSFMT_N; i++) { |
298 |
do_recursion(&array[i], &array[i - DSFMT_N], |
299 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
300 |
convert_c0o1(&array[i - DSFMT_N]); |
301 |
} |
302 |
for (j = 0; j < 2 * DSFMT_N - size; j++) { |
303 |
dsfmt->status[j] = array[j + size - DSFMT_N]; |
304 |
} |
305 |
for (; i < size; i++, j++) { |
306 |
do_recursion(&array[i], &array[i - DSFMT_N], |
307 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
308 |
dsfmt->status[j] = array[i]; |
309 |
convert_c0o1(&array[i - DSFMT_N]); |
310 |
} |
311 |
for (i = size - DSFMT_N; i < size; i++) { |
312 |
convert_c0o1(&array[i]); |
313 |
} |
314 |
dsfmt->status[DSFMT_N] = lung; |
315 |
} |
316 |
|
317 |
/** |
318 |
* This function fills the user-specified array with double precision |
319 |
* floating point pseudorandom numbers of the IEEE 754 format. |
320 |
* @param dsfmt dsfmt state vector. |
321 |
* @param array an 128-bit array to be filled by pseudorandom numbers. |
322 |
* @param size number of 128-bit pseudorandom numbers to be generated. |
323 |
*/ |
324 |
inline static void gen_rand_array_o0o1(dsfmt_t *dsfmt, w128_t *array, |
325 |
int size) { |
326 |
int i, j; |
327 |
w128_t lung; |
328 |
|
329 |
lung = dsfmt->status[DSFMT_N]; |
330 |
do_recursion(&array[0], &dsfmt->status[0], &dsfmt->status[DSFMT_POS1], |
331 |
&lung); |
332 |
for (i = 1; i < DSFMT_N - DSFMT_POS1; i++) { |
333 |
do_recursion(&array[i], &dsfmt->status[i], |
334 |
&dsfmt->status[i + DSFMT_POS1], &lung); |
335 |
} |
336 |
for (; i < DSFMT_N; i++) { |
337 |
do_recursion(&array[i], &dsfmt->status[i], |
338 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
339 |
} |
340 |
for (; i < size - DSFMT_N; i++) { |
341 |
do_recursion(&array[i], &array[i - DSFMT_N], |
342 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
343 |
convert_o0o1(&array[i - DSFMT_N]); |
344 |
} |
345 |
for (j = 0; j < 2 * DSFMT_N - size; j++) { |
346 |
dsfmt->status[j] = array[j + size - DSFMT_N]; |
347 |
} |
348 |
for (; i < size; i++, j++) { |
349 |
do_recursion(&array[i], &array[i - DSFMT_N], |
350 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
351 |
dsfmt->status[j] = array[i]; |
352 |
convert_o0o1(&array[i - DSFMT_N]); |
353 |
} |
354 |
for (i = size - DSFMT_N; i < size; i++) { |
355 |
convert_o0o1(&array[i]); |
356 |
} |
357 |
dsfmt->status[DSFMT_N] = lung; |
358 |
} |
359 |
|
360 |
/** |
361 |
* This function fills the user-specified array with double precision |
362 |
* floating point pseudorandom numbers of the IEEE 754 format. |
363 |
* @param dsfmt dsfmt state vector. |
364 |
* @param array an 128-bit array to be filled by pseudorandom numbers. |
365 |
* @param size number of 128-bit pseudorandom numbers to be generated. |
366 |
*/ |
367 |
inline static void gen_rand_array_o0c1(dsfmt_t *dsfmt, w128_t *array, |
368 |
int size) { |
369 |
int i, j; |
370 |
w128_t lung; |
371 |
|
372 |
lung = dsfmt->status[DSFMT_N]; |
373 |
do_recursion(&array[0], &dsfmt->status[0], &dsfmt->status[DSFMT_POS1], |
374 |
&lung); |
375 |
for (i = 1; i < DSFMT_N - DSFMT_POS1; i++) { |
376 |
do_recursion(&array[i], &dsfmt->status[i], |
377 |
&dsfmt->status[i + DSFMT_POS1], &lung); |
378 |
} |
379 |
for (; i < DSFMT_N; i++) { |
380 |
do_recursion(&array[i], &dsfmt->status[i], |
381 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
382 |
} |
383 |
for (; i < size - DSFMT_N; i++) { |
384 |
do_recursion(&array[i], &array[i - DSFMT_N], |
385 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
386 |
convert_o0c1(&array[i - DSFMT_N]); |
387 |
} |
388 |
for (j = 0; j < 2 * DSFMT_N - size; j++) { |
389 |
dsfmt->status[j] = array[j + size - DSFMT_N]; |
390 |
} |
391 |
for (; i < size; i++, j++) { |
392 |
do_recursion(&array[i], &array[i - DSFMT_N], |
393 |
&array[i + DSFMT_POS1 - DSFMT_N], &lung); |
394 |
dsfmt->status[j] = array[i]; |
395 |
convert_o0c1(&array[i - DSFMT_N]); |
396 |
} |
397 |
for (i = size - DSFMT_N; i < size; i++) { |
398 |
convert_o0c1(&array[i]); |
399 |
} |
400 |
dsfmt->status[DSFMT_N] = lung; |
401 |
} |
402 |
|
403 |
/** |
404 |
* This function represents a function used in the initialization |
405 |
* by init_by_array |
406 |
* @param x 32-bit integer |
407 |
* @return 32-bit integer |
408 |
*/ |
409 |
static uint32_t ini_func1(uint32_t x) { |
410 |
return (x ^ (x >> 27)) * (uint32_t)1664525UL; |
411 |
} |
412 |
|
413 |
/** |
414 |
* This function represents a function used in the initialization |
415 |
* by init_by_array |
416 |
* @param x 32-bit integer |
417 |
* @return 32-bit integer |
418 |
*/ |
419 |
static uint32_t ini_func2(uint32_t x) { |
420 |
return (x ^ (x >> 27)) * (uint32_t)1566083941UL; |
421 |
} |
422 |
|
423 |
/** |
424 |
* This function initializes the internal state array to fit the IEEE |
425 |
* 754 format. |
426 |
* @param dsfmt dsfmt state vector. |
427 |
*/ |
428 |
static void initial_mask(dsfmt_t *dsfmt) { |
429 |
int i; |
430 |
uint64_t *psfmt; |
431 |
|
432 |
psfmt = &dsfmt->status[0].u[0]; |
433 |
for (i = 0; i < DSFMT_N * 2; i++) { |
434 |
psfmt[i] = (psfmt[i] & DSFMT_LOW_MASK) | DSFMT_HIGH_CONST; |
435 |
} |
436 |
} |
437 |
|
438 |
/** |
439 |
* This function certificate the period of 2^{SFMT_MEXP}-1. |
440 |
* @param dsfmt dsfmt state vector. |
441 |
*/ |
442 |
static void period_certification(dsfmt_t *dsfmt) { |
443 |
uint64_t pcv[2] = {DSFMT_PCV1, DSFMT_PCV2}; |
444 |
uint64_t tmp[2]; |
445 |
uint64_t inner; |
446 |
int i; |
447 |
#if (DSFMT_PCV2 & 1) != 1 |
448 |
int j; |
449 |
uint64_t work; |
450 |
#endif |
451 |
|
452 |
tmp[0] = (dsfmt->status[DSFMT_N].u[0] ^ DSFMT_FIX1); |
453 |
tmp[1] = (dsfmt->status[DSFMT_N].u[1] ^ DSFMT_FIX2); |
454 |
|
455 |
inner = tmp[0] & pcv[0]; |
456 |
inner ^= tmp[1] & pcv[1]; |
457 |
for (i = 32; i > 0; i >>= 1) { |
458 |
inner ^= inner >> i; |
459 |
} |
460 |
inner &= 1; |
461 |
/* check OK */ |
462 |
if (inner == 1) { |
463 |
return; |
464 |
} |
465 |
/* check NG, and modification */ |
466 |
#if (DSFMT_PCV2 & 1) == 1 |
467 |
dsfmt->status[DSFMT_N].u[1] ^= 1; |
468 |
#else |
469 |
for (i = 1; i >= 0; i--) { |
470 |
work = 1; |
471 |
for (j = 0; j < 64; j++) { |
472 |
if ((work & pcv[i]) != 0) { |
473 |
dsfmt->status[DSFMT_N].u[i] ^= work; |
474 |
return; |
475 |
} |
476 |
work = work << 1; |
477 |
} |
478 |
} |
479 |
#endif |
480 |
return; |
481 |
} |
482 |
|
483 |
/*---------------- |
484 |
PUBLIC FUNCTIONS |
485 |
----------------*/ |
486 |
/** |
487 |
* This function returns the identification string. The string shows |
488 |
* the Mersenne exponent, and all parameters of this generator. |
489 |
* @return id string. |
490 |
*/ |
491 |
const char *dsfmt_get_idstring(void) { |
492 |
return DSFMT_IDSTR; |
493 |
} |
494 |
|
495 |
/** |
496 |
* This function returns the minimum size of array used for \b |
497 |
* fill_array functions. |
498 |
* @return minimum size of array used for fill_array functions. |
499 |
*/ |
500 |
int dsfmt_get_min_array_size(void) { |
501 |
return DSFMT_N64; |
502 |
} |
503 |
|
504 |
/** |
505 |
* This function fills the internal state array with double precision |
506 |
* floating point pseudorandom numbers of the IEEE 754 format. |
507 |
* @param dsfmt dsfmt state vector. |
508 |
*/ |
509 |
void dsfmt_gen_rand_all(dsfmt_t *dsfmt) { |
510 |
int i; |
511 |
w128_t lung; |
512 |
|
513 |
lung = dsfmt->status[DSFMT_N]; |
514 |
do_recursion(&dsfmt->status[0], &dsfmt->status[0], |
515 |
&dsfmt->status[DSFMT_POS1], &lung); |
516 |
for (i = 1; i < DSFMT_N - DSFMT_POS1; i++) { |
517 |
do_recursion(&dsfmt->status[i], &dsfmt->status[i], |
518 |
&dsfmt->status[i + DSFMT_POS1], &lung); |
519 |
} |
520 |
for (; i < DSFMT_N; i++) { |
521 |
do_recursion(&dsfmt->status[i], &dsfmt->status[i], |
522 |
&dsfmt->status[i + DSFMT_POS1 - DSFMT_N], &lung); |
523 |
} |
524 |
dsfmt->status[DSFMT_N] = lung; |
525 |
} |
526 |
|
527 |
/** |
528 |
* This function generates double precision floating point |
529 |
* pseudorandom numbers which distribute in the range [1, 2) to the |
530 |
* specified array[] by one call. The number of pseudorandom numbers |
531 |
* is specified by the argument \b size, which must be at least (SFMT_MEXP |
532 |
* / 128) * 2 and a multiple of two. The function |
533 |
* get_min_array_size() returns this minimum size. The generation by |
534 |
* this function is much faster than the following fill_array_xxx functions. |
535 |
* |
536 |
* For initialization, init_gen_rand() or init_by_array() must be called |
537 |
* before the first call of this function. This function can not be |
538 |
* used after calling genrand_xxx functions, without initialization. |
539 |
* |
540 |
* @param dsfmt dsfmt state vector. |
541 |
* @param array an array where pseudorandom numbers are filled |
542 |
* by this function. The pointer to the array must be "aligned" |
543 |
* (namely, must be a multiple of 16) in the SIMD version, since it |
544 |
* refers to the address of a 128-bit integer. In the standard C |
545 |
* version, the pointer is arbitrary. |
546 |
* |
547 |
* @param size the number of 64-bit pseudorandom integers to be |
548 |
* generated. size must be a multiple of 2, and greater than or equal |
549 |
* to (SFMT_MEXP / 128) * 2. |
550 |
* |
551 |
* @note \b memalign or \b posix_memalign is available to get aligned |
552 |
* memory. Mac OSX doesn't have these functions, but \b malloc of OSX |
553 |
* returns the pointer to the aligned memory block. |
554 |
*/ |
555 |
void dsfmt_fill_array_close1_open2(dsfmt_t *dsfmt, double array[], int size) { |
556 |
assert(size % 2 == 0); |
557 |
assert(size >= DSFMT_N64); |
558 |
gen_rand_array_c1o2(dsfmt, (w128_t *)array, size / 2); |
559 |
} |
560 |
|
561 |
/** |
562 |
* This function generates double precision floating point |
563 |
* pseudorandom numbers which distribute in the range (0, 1] to the |
564 |
* specified array[] by one call. This function is the same as |
565 |
* fill_array_close1_open2() except the distribution range. |
566 |
* |
567 |
* @param dsfmt dsfmt state vector. |
568 |
* @param array an array where pseudorandom numbers are filled |
569 |
* by this function. |
570 |
* @param size the number of pseudorandom numbers to be generated. |
571 |
* see also \sa fill_array_close1_open2() |
572 |
*/ |
573 |
void dsfmt_fill_array_open_close(dsfmt_t *dsfmt, double array[], int size) { |
574 |
assert(size % 2 == 0); |
575 |
assert(size >= DSFMT_N64); |
576 |
gen_rand_array_o0c1(dsfmt, (w128_t *)array, size / 2); |
577 |
} |
578 |
|
579 |
/** |
580 |
* This function generates double precision floating point |
581 |
* pseudorandom numbers which distribute in the range [0, 1) to the |
582 |
* specified array[] by one call. This function is the same as |
583 |
* fill_array_close1_open2() except the distribution range. |
584 |
* |
585 |
* @param array an array where pseudorandom numbers are filled |
586 |
* by this function. |
587 |
* @param dsfmt dsfmt state vector. |
588 |
* @param size the number of pseudorandom numbers to be generated. |
589 |
* see also \sa fill_array_close1_open2() |
590 |
*/ |
591 |
void dsfmt_fill_array_close_open(dsfmt_t *dsfmt, double array[], int size) { |
592 |
assert(size % 2 == 0); |
593 |
assert(size >= DSFMT_N64); |
594 |
gen_rand_array_c0o1(dsfmt, (w128_t *)array, size / 2); |
595 |
} |
596 |
|
597 |
/** |
598 |
* This function generates double precision floating point |
599 |
* pseudorandom numbers which distribute in the range (0, 1) to the |
600 |
* specified array[] by one call. This function is the same as |
601 |
* fill_array_close1_open2() except the distribution range. |
602 |
* |
603 |
* @param dsfmt dsfmt state vector. |
604 |
* @param array an array where pseudorandom numbers are filled |
605 |
* by this function. |
606 |
* @param size the number of pseudorandom numbers to be generated. |
607 |
* see also \sa fill_array_close1_open2() |
608 |
*/ |
609 |
void dsfmt_fill_array_open_open(dsfmt_t *dsfmt, double array[], int size) { |
610 |
assert(size % 2 == 0); |
611 |
assert(size >= DSFMT_N64); |
612 |
gen_rand_array_o0o1(dsfmt, (w128_t *)array, size / 2); |
613 |
} |
614 |
|
615 |
#if defined(__INTEL_COMPILER) |
616 |
# pragma warning(disable:981) |
617 |
#endif |
618 |
/** |
619 |
* This function initializes the internal state array with a 32-bit |
620 |
* integer seed. |
621 |
* @param dsfmt dsfmt state vector. |
622 |
* @param seed a 32-bit integer used as the seed. |
623 |
* @param mexp caller's mersenne expornent |
624 |
*/ |
625 |
void dsfmt_chk_init_gen_rand(dsfmt_t *dsfmt, uint32_t seed, int mexp) { |
626 |
int i; |
627 |
uint32_t *psfmt; |
628 |
|
629 |
/* make sure caller program is compiled with the same MEXP */ |
630 |
if (mexp != dsfmt_mexp) { |
631 |
fprintf(stderr, "DSFMT_MEXP doesn't match with dSFMT.c\n"); |
632 |
exit(1); |
633 |
} |
634 |
psfmt = &dsfmt->status[0].u32[0]; |
635 |
psfmt[idxof(0)] = seed; |
636 |
for (i = 1; i < (DSFMT_N + 1) * 4; i++) { |
637 |
psfmt[idxof(i)] = 1812433253UL |
638 |
* (psfmt[idxof(i - 1)] ^ (psfmt[idxof(i - 1)] >> 30)) + i; |
639 |
} |
640 |
initial_mask(dsfmt); |
641 |
period_certification(dsfmt); |
642 |
dsfmt->idx = DSFMT_N64; |
643 |
#if defined(HAVE_SSE2) |
644 |
setup_const(); |
645 |
#endif |
646 |
} |
647 |
|
648 |
/** |
649 |
* This function initializes the internal state array, |
650 |
* with an array of 32-bit integers used as the seeds |
651 |
* @param dsfmt dsfmt state vector. |
652 |
* @param init_key the array of 32-bit integers, used as a seed. |
653 |
* @param key_length the length of init_key. |
654 |
* @param mexp caller's mersenne expornent |
655 |
*/ |
656 |
void dsfmt_chk_init_by_array(dsfmt_t *dsfmt, uint32_t init_key[], |
657 |
int key_length, int mexp) { |
658 |
int i, j, count; |
659 |
uint32_t r; |
660 |
uint32_t *psfmt32; |
661 |
int lag; |
662 |
int mid; |
663 |
int size = (DSFMT_N + 1) * 4; /* pulmonary */ |
664 |
|
665 |
/* make sure caller program is compiled with the same MEXP */ |
666 |
if (mexp != dsfmt_mexp) { |
667 |
fprintf(stderr, "DSFMT_MEXP doesn't match with dSFMT.c\n"); |
668 |
exit(1); |
669 |
} |
670 |
if (size >= 623) { |
671 |
lag = 11; |
672 |
} else if (size >= 68) { |
673 |
lag = 7; |
674 |
} else if (size >= 39) { |
675 |
lag = 5; |
676 |
} else { |
677 |
lag = 3; |
678 |
} |
679 |
mid = (size - lag) / 2; |
680 |
|
681 |
psfmt32 = &dsfmt->status[0].u32[0]; |
682 |
memset(dsfmt->status, 0x8b, sizeof(dsfmt->status)); |
683 |
if (key_length + 1 > size) { |
684 |
count = key_length + 1; |
685 |
} else { |
686 |
count = size; |
687 |
} |
688 |
r = ini_func1(psfmt32[idxof(0)] ^ psfmt32[idxof(mid % size)] |
689 |
^ psfmt32[idxof((size - 1) % size)]); |
690 |
psfmt32[idxof(mid % size)] += r; |
691 |
r += key_length; |
692 |
psfmt32[idxof((mid + lag) % size)] += r; |
693 |
psfmt32[idxof(0)] = r; |
694 |
count--; |
695 |
for (i = 1, j = 0; (j < count) && (j < key_length); j++) { |
696 |
r = ini_func1(psfmt32[idxof(i)] |
697 |
^ psfmt32[idxof((i + mid) % size)] |
698 |
^ psfmt32[idxof((i + size - 1) % size)]); |
699 |
psfmt32[idxof((i + mid) % size)] += r; |
700 |
r += init_key[j] + i; |
701 |
psfmt32[idxof((i + mid + lag) % size)] += r; |
702 |
psfmt32[idxof(i)] = r; |
703 |
i = (i + 1) % size; |
704 |
} |
705 |
for (; j < count; j++) { |
706 |
r = ini_func1(psfmt32[idxof(i)] |
707 |
^ psfmt32[idxof((i + mid) % size)] |
708 |
^ psfmt32[idxof((i + size - 1) % size)]); |
709 |
psfmt32[idxof((i + mid) % size)] += r; |
710 |
r += i; |
711 |
psfmt32[idxof((i + mid + lag) % size)] += r; |
712 |
psfmt32[idxof(i)] = r; |
713 |
i = (i + 1) % size; |
714 |
} |
715 |
for (j = 0; j < size; j++) { |
716 |
r = ini_func2(psfmt32[idxof(i)] |
717 |
+ psfmt32[idxof((i + mid) % size)] |
718 |
+ psfmt32[idxof((i + size - 1) % size)]); |
719 |
psfmt32[idxof((i + mid) % size)] ^= r; |
720 |
r -= i; |
721 |
psfmt32[idxof((i + mid + lag) % size)] ^= r; |
722 |
psfmt32[idxof(i)] = r; |
723 |
i = (i + 1) % size; |
724 |
} |
725 |
initial_mask(dsfmt); |
726 |
period_certification(dsfmt); |
727 |
dsfmt->idx = DSFMT_N64; |
728 |
#if defined(HAVE_SSE2) |
729 |
setup_const(); |
730 |
#endif |
731 |
} |
732 |
#if defined(__INTEL_COMPILER) |
733 |
# pragma warning(default:981) |
734 |
#endif |