OSchip
/
OpenBLAS
Not watched
1
0
Fork 0
Code
Releases 66 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
Tree: bfaaa975e6
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'bfaaa975e6'
${ noResults }
OpenBLAS/kernel/x86/asum_sse2.S

319 lines
6.7 kB
Raw Blame History 

  1. /*********************************************************************/

  2. /* Copyright 2009, 2010 The University of Texas at Austin.           */

  3. /* All rights reserved.                                              */

  4. /*                                                                   */

  5. /* Redistribution and use in source and binary forms, with or        */

  6. /* without modification, are permitted provided that the following   */

  7. /* conditions are met:                                               */

  8. /*                                                                   */

  9. /*   1. Redistributions of source code must retain the above         */

  10. /*      copyright notice, this list of conditions and the following  */

  11. /*      disclaimer.                                                  */

  12. /*                                                                   */

  13. /*   2. Redistributions in binary form must reproduce the above      */

  14. /*      copyright notice, this list of conditions and the following  */

  15. /*      disclaimer in the documentation and/or other materials       */

  16. /*      provided with the distribution.                              */

  17. /*                                                                   */

  18. /*    THIS  SOFTWARE IS PROVIDED  BY THE  UNIVERSITY OF  TEXAS AT    */

  19. /*    AUSTIN  ``AS IS''  AND ANY  EXPRESS OR  IMPLIED WARRANTIES,    */

  20. /*    INCLUDING, BUT  NOT LIMITED  TO, THE IMPLIED  WARRANTIES OF    */

  21. /*    MERCHANTABILITY  AND FITNESS FOR  A PARTICULAR  PURPOSE ARE    */

  22. /*    DISCLAIMED.  IN  NO EVENT SHALL THE UNIVERSITY  OF TEXAS AT    */

  23. /*    AUSTIN OR CONTRIBUTORS BE  LIABLE FOR ANY DIRECT, INDIRECT,    */

  24. /*    INCIDENTAL,  SPECIAL, EXEMPLARY,  OR  CONSEQUENTIAL DAMAGES    */

  25. /*    (INCLUDING, BUT  NOT LIMITED TO,  PROCUREMENT OF SUBSTITUTE    */

  26. /*    GOODS  OR  SERVICES; LOSS  OF  USE,  DATA,  OR PROFITS;  OR    */

  27. /*    BUSINESS INTERRUPTION) HOWEVER CAUSED  AND ON ANY THEORY OF    */

  28. /*    LIABILITY, WHETHER  IN CONTRACT, STRICT  LIABILITY, OR TORT    */

  29. /*    (INCLUDING NEGLIGENCE OR OTHERWISE)  ARISING IN ANY WAY OUT    */

  30. /*    OF  THE  USE OF  THIS  SOFTWARE,  EVEN  IF ADVISED  OF  THE    */

  31. /*    POSSIBILITY OF SUCH DAMAGE.                                    */

  32. /*                                                                   */

  33. /* The views and conclusions contained in the software and           */

  34. /* documentation are those of the authors and should not be          */

  35. /* interpreted as representing official policies, either expressed   */

  36. /* or implied, of The University of Texas at Austin.                 */

  37. /*********************************************************************/

  38. 

  39. #define ASSEMBLER

  40. #include "common.h"

  41. 	

  42. #define STACK	 8

  43. #define ARGS     0

  44. 

  45. #define STACK_M		 4 + STACK + ARGS(%esp)

  46. #define STACK_X		 8 + STACK + ARGS(%esp)

  47. #define STACK_INCX	12 + STACK + ARGS(%esp)

  48. 

  49. #define I	%eax

  50. #define M	%ecx

  51. #define X	%esi

  52. #define INCX	%ebx

  53. 

  54. #include "l1param.h"

  55. 

  56. 	PROLOGUE

  57. 	PROFCODE

  58. 

  59. 	pushl	%esi

  60. 	pushl	%ebx

  61. 

  62. 	movl	STACK_M, M

  63. 	movl	STACK_X, X

  64. 	movl	STACK_INCX, INCX

  65. 

  66. 	xorps	%xmm0, %xmm0

  67. 	xorps	%xmm1, %xmm1

  68. 

  69. 	testl	M, M

  70. 	jle	.L999

  71. 	testl	INCX, INCX

  72. 	jle	.L999

  73. 

  74. 	pcmpeqb	%xmm3, %xmm3

  75. 	psrlq	$1,    %xmm3

  76. 

  77. 	sall	$BASE_SHIFT, INCX

  78. 

  79. 	subl	$-16 * SIZE, X

  80. 

  81. 	cmpl	$SIZE, INCX

  82. 	jne	.L40

  83. 

  84. 	testl	$SIZE, X

  85. 	je	.L05

  86. 

  87. 	movsd	-16 * SIZE(X), %xmm0

  88. 	addl	$SIZE, X

  89. 

  90. 	andps	%xmm3, %xmm0

  91. 	subl	$1, M

  92. 	jle	.L999

  93. 	ALIGN_3

  94. 

  95. .L05:

  96. 	movl	M,  I

  97. 	sarl	$4, I

  98. 	jle	.L20

  99. 

  100. 	movaps	-16 * SIZE(X), %xmm4

  101. 	movaps	-14 * SIZE(X), %xmm5

  102. 	movaps	-12 * SIZE(X), %xmm6

  103. 	movaps	-10 * SIZE(X), %xmm7

  104. 

  105. 	decl	I

  106. 	jle	.L11

  107. 	ALIGN_4

  108. 	

  109. .L10:

  110. #ifdef PREFETCH

  111. 	PREFETCH (PREFETCHSIZE +  0) - PREOFFSET(X)

  112. #endif

  113. 

  114. 	andps	%xmm3, %xmm4

  115. 	addpd	%xmm4, %xmm0

  116. 	movaps	 -8 * SIZE(X), %xmm4

  117. 

  118. 	andps	%xmm3, %xmm5

  119. 	addpd	%xmm5, %xmm1

  120. 	movaps	 -6 * SIZE(X), %xmm5

  121. 

  122. 	andps	%xmm3, %xmm6

  123. 	addpd	%xmm6, %xmm0

  124. 	movaps	 -4 * SIZE(X), %xmm6

  125. 

  126. 	andps	%xmm3, %xmm7

  127. 	addpd	%xmm7, %xmm1

  128. 	movaps	 -2 * SIZE(X), %xmm7

  129. 

  130. #if defined(PREFETCH) && !defined(FETCH128)

  131. 	PREFETCH (PREFETCHSIZE +  64) - PREOFFSET(X)

  132. #endif

  133. 

  134. 	andps	%xmm3, %xmm4

  135. 	addpd	%xmm4, %xmm0

  136. 	movaps	  0 * SIZE(X), %xmm4

  137. 

  138. 	andps	%xmm3, %xmm5

  139. 	addpd	%xmm5, %xmm1

  140. 	movaps	  2 * SIZE(X), %xmm5

  141. 

  142. 	andps	%xmm3, %xmm6

  143. 	addpd	%xmm6, %xmm0

  144. 	movaps	  4 * SIZE(X), %xmm6

  145. 

  146. 	andps	%xmm3, %xmm7

  147. 	addpd	%xmm7, %xmm1

  148. 	movaps	  6 * SIZE(X), %xmm7

  149. 

  150. 	subl	$-16 * SIZE, X

  151. 	decl	I

  152. 	jg	.L10

  153. 	ALIGN_4

  154. 

  155. .L11:

  156. 	andps	%xmm3, %xmm4

  157. 	addpd	%xmm4, %xmm0

  158. 	movaps	 -8 * SIZE(X), %xmm4

  159. 

  160. 	andps	%xmm3, %xmm5

  161. 	addpd	%xmm5, %xmm1

  162. 	movaps	 -6 * SIZE(X), %xmm5

  163. 

  164. 	andps	%xmm3, %xmm6

  165. 	addpd	%xmm6, %xmm0

  166. 	movaps	 -4 * SIZE(X), %xmm6

  167. 

  168. 	andps	%xmm3, %xmm7

  169. 	addpd	%xmm7, %xmm1

  170. 	movaps	 -2 * SIZE(X), %xmm7

  171. 

  172. 	andps	%xmm3, %xmm4

  173. 	addpd	%xmm4, %xmm0

  174. 	andps	%xmm3, %xmm5

  175. 	addpd	%xmm5, %xmm1

  176. 	andps	%xmm3, %xmm6

  177. 	addpd	%xmm6, %xmm0

  178. 	andps	%xmm3, %xmm7

  179. 	addpd	%xmm7, %xmm1

  180. 

  181. 	subl	$-16 * SIZE, X

  182. 	ALIGN_3

  183. 

  184. .L20:

  185. 	andl	$15,  M

  186. 	jle	.L999

  187. 

  188. 	testl	$8, M

  189. 	je	.L21

  190. 

  191. 	movaps	-16 * SIZE(X), %xmm4

  192. 	movaps	-14 * SIZE(X), %xmm5

  193. 	movaps	-12 * SIZE(X), %xmm6

  194. 	movaps	-10 * SIZE(X), %xmm7

  195. 

  196. 	andps	%xmm3, %xmm4

  197. 	addpd	%xmm4, %xmm0

  198. 	andps	%xmm3, %xmm5

  199. 	addpd	%xmm5, %xmm1

  200. 	andps	%xmm3, %xmm6

  201. 	addpd	%xmm6, %xmm0

  202. 	andps	%xmm3, %xmm7

  203. 	addpd	%xmm7, %xmm1

  204. 	addl	$8 * SIZE, X

  205. 	ALIGN_3

  206. 

  207. .L21:

  208. 	testl	$4, M

  209. 	je	.L22

  210. 

  211. 	movaps	-16 * SIZE(X), %xmm4

  212. 	movaps	-14 * SIZE(X), %xmm5

  213. 

  214. 	andps	%xmm3, %xmm4

  215. 	addpd	%xmm4, %xmm0

  216. 	andps	%xmm3, %xmm5

  217. 	addpd	%xmm5, %xmm1

  218. 

  219. 	addl	$4 * SIZE, X

  220. 	ALIGN_3	

  221. 

  222. .L22:

  223. 	testl	$2, M

  224. 	je	.L23

  225. 

  226. 	movaps	-16 * SIZE(X), %xmm4

  227. 	andps	%xmm3, %xmm4

  228. 	addpd	%xmm4, %xmm0

  229. 	addl	$2 * SIZE, X

  230. 	

  231. .L23:

  232. 	testl	$1, M

  233. 	je	.L999

  234. 

  235. #ifdef movsd

  236. 	xorps	%xmm4, %xmm4

  237. #endif

  238. 	movsd	-16 * SIZE(X), %xmm4

  239. 	andps	%xmm3, %xmm4

  240. 	addsd	%xmm4, %xmm1

  241. 	jmp	.L999

  242. 	ALIGN_3

  243. 

  244. .L40:

  245. 	movl	M,  I

  246. 	sarl	$3, I

  247. 	jle	.L60

  248. 	ALIGN_4

  249. 	

  250. .L50:

  251. 	movsd	-16 * SIZE(X), %xmm4

  252. 	addl	INCX, X

  253. 	movhps	-16 * SIZE(X), %xmm4

  254. 	addl	INCX, X

  255. 	andps	%xmm3, %xmm4

  256. 	addpd	%xmm4, %xmm0

  257. 

  258. 	movsd	-16 * SIZE(X), %xmm5

  259. 	addl	INCX, X

  260. 	movhps	-16 * SIZE(X), %xmm5

  261. 	addl	INCX, X

  262. 	andps	%xmm3, %xmm5

  263. 	addpd	%xmm5, %xmm1

  264. 

  265. 	movsd	-16 * SIZE(X), %xmm6

  266. 	addl	INCX, X

  267. 	movhps	-16 * SIZE(X), %xmm6

  268. 	addl	INCX, X

  269. 	andps	%xmm3, %xmm6

  270. 	addpd	%xmm6, %xmm0

  271. 

  272. 	movsd	-16 * SIZE(X), %xmm7

  273. 	addl	INCX, X

  274. 	movhps	-16 * SIZE(X), %xmm7

  275. 	addl	INCX, X

  276. 	andps	%xmm3, %xmm7

  277. 	addpd	%xmm7, %xmm1

  278. 

  279. 	decl	I

  280. 	jg	.L50

  281. 	ALIGN_4

  282. 

  283. .L60:

  284. #ifdef movsd

  285. 	xorps	%xmm4, %xmm4

  286. #endif

  287. 	andl	$7,  M

  288. 	jle	.L999

  289. 	ALIGN_4

  290. 

  291. .L61:

  292. 	movsd	-16 * SIZE(X), %xmm4

  293. 	andps	%xmm3, %xmm4

  294. 	addsd	%xmm4, %xmm0

  295. 	addl	INCX, X

  296. 	decl	M

  297. 	jg	.L61

  298. 	ALIGN_4

  299. 

  300. .L999:

  301. 	addpd	%xmm1, %xmm0

  302. 

  303. #ifndef HAVE_SSE3

  304. 	movaps	%xmm0,  %xmm1

  305. 	unpckhpd %xmm0, %xmm0

  306. 	addsd	 %xmm1, %xmm0

  307. #else

  308. 	haddpd	%xmm0, %xmm0

  309. #endif

  310. 

  311. 	movsd	%xmm0, STACK_M

  312. 	fldl	STACK_M

  313. 	popl	%ebx

  314. 	popl	%esi

  315. 	ret

  316. 

  317. 	EPILOGUE