Debian Codegraph @

libmp3lame/i386/scalar.nas

raw
1 ;
2 ; (C) Frank Klemm 1995,99,2000
3 ; Dedicated to the LAME project
4 ;
5 ;
6 %include "nasm.h"
7
8 segment_code
9
10 ; float_t scalar04_float32_i387 (
11 ; const float32_t* const p,
12 ; const float32_t* const q );
13
14 proc scalar04_float32_i387
15 %$p arg 4
16 %$q arg 4
17 ;;; alloc
18
19 mov eax,[sp(%$p)]
20 mov edx,[sp(%$q)]
21 fld dword [eax]
22 fmul dword [edx]
23 fld dword [eax + 4]
24 fmul dword [edx + 4]
25 faddp st1,st0
26 fld dword [eax + 8]
27 fmul dword [edx + 8]
28 faddp st1,st0
29 fld dword [eax + 12]
30 fmul dword [edx + 12]
31 faddp st1,st0
32 endproc
33
34
35 proc scalar08_float32_i387
36 %$p arg 4
37 %$q arg 4
38 ;;; alloc
39
40 mov eax,[sp(%$p)]
41 mov edx,[sp(%$q)]
42 fld dword [eax]
43 fmul dword [edx]
44 fld dword [eax + 4]
45 fmul dword [edx + 4]
46 faddp st1,st0
47 fld dword [eax + 8]
48 fmul dword [edx + 8]
49 faddp st1,st0
50 fld dword [eax + 12]
51 fmul dword [edx + 12]
52 faddp st1,st0
53 fld dword [eax + 16]
54 fmul dword [edx + 16]
55 faddp st1,st0
56 fld dword [eax + 20]
57 fmul dword [edx + 20]
58 faddp st1,st0
59 fld dword [eax + 24]
60 fmul dword [edx + 24]
61 faddp st1,st0
62 fld dword [eax + 28]
63 fmul dword [edx + 28]
64 faddp st1,st0
65 endproc
66
67
68 proc scalar12_float32_i387
69 %$p arg 4
70 %$q arg 4
71 ;;; alloc
72
73 mov eax,[sp(%$p)]
74 mov edx,[sp(%$q)]
75 fld dword [eax]
76 fmul dword [edx]
77 fld dword [eax + 4]
78 fmul dword [edx + 4]
79 faddp st1,st0
80 fld dword [eax + 8]
81 fmul dword [edx + 8]
82 faddp st1,st0
83 fld dword [eax + 12]
84 fmul dword [edx + 12]
85 faddp st1,st0
86 fld dword [eax + 16]
87 fmul dword [edx + 16]
88 faddp st1,st0
89 fld dword [eax + 20]
90 fmul dword [edx + 20]
91 faddp st1,st0
92 fld dword [eax + 24]
93 fmul dword [edx + 24]
94 faddp st1,st0
95 fld dword [eax + 28]
96 fmul dword [edx + 28]
97 faddp st1,st0
98 fld dword [eax + 32]
99 fmul dword [edx + 32]
100 faddp st1,st0
101 fld dword [eax + 36]
102 fmul dword [edx + 36]
103 faddp st1,st0
104 fld dword [eax + 40]
105 fmul dword [edx + 40]
106 faddp st1,st0
107 fld dword [eax + 44]
108 fmul dword [edx + 44]
109 faddp st1,st0
110 endproc
111
112
113 proc scalar16_float32_i387
114 %$p arg 4
115 %$q arg 4
116 ;;; alloc
117
118 mov eax,[sp(%$p)]
119 mov edx,[sp(%$q)]
120 fld dword [eax]
121 fmul dword [edx]
122 fld dword [eax + 4]
123 fmul dword [edx + 4]
124 faddp st1,st0
125 fld dword [eax + 8]
126 fmul dword [edx + 8]
127 faddp st1,st0
128 fld dword [eax + 12]
129 fmul dword [edx + 12]
130 faddp st1,st0
131 fld dword [eax + 16]
132 fmul dword [edx + 16]
133 faddp st1,st0
134 fld dword [eax + 20]
135 fmul dword [edx + 20]
136 faddp st1,st0
137 fld dword [eax + 24]
138 fmul dword [edx + 24]
139 faddp st1,st0
140 fld dword [eax + 28]
141 fmul dword [edx + 28]
142 faddp st1,st0
143 fld dword [eax + 32]
144 fmul dword [edx + 32]
145 faddp st1,st0
146 fld dword [eax + 36]
147 fmul dword [edx + 36]
148 faddp st1,st0
149 fld dword [eax + 40]
150 fmul dword [edx + 40]
151 faddp st1,st0
152 fld dword [eax + 44]
153 fmul dword [edx + 44]
154 faddp st1,st0
155 fld dword [eax + 48]
156 fmul dword [edx + 48]
157 faddp st1,st0
158 fld dword [eax + 52]
159 fmul dword [edx + 52]
160 faddp st1,st0
161 fld dword [eax + 56]
162 fmul dword [edx + 56]
163 faddp st1,st0
164 fld dword [eax + 60]
165 fmul dword [edx + 60]
166 faddp st1,st0
167 endproc
168
169
170 proc scalar20_float32_i387
171 %$p arg 4
172 %$q arg 4
173 ;;; alloc
174
175 mov eax,[sp(%$p)]
176 mov edx,[sp(%$q)]
177 fld dword [eax]
178 fmul dword [edx]
179 fld dword [eax + 4]
180 fmul dword [edx + 4]
181 faddp st1,st0
182 fld dword [eax + 8]
183 fmul dword [edx + 8]
184 faddp st1,st0
185 fld dword [eax + 12]
186 fmul dword [edx + 12]
187 faddp st1,st0
188 fld dword [eax + 16]
189 fmul dword [edx + 16]
190 faddp st1,st0
191 fld dword [eax + 20]
192 fmul dword [edx + 20]
193 faddp st1,st0
194 fld dword [eax + 24]
195 fmul dword [edx + 24]
196 faddp st1,st0
197 fld dword [eax + 28]
198 fmul dword [edx + 28]
199 faddp st1,st0
200 fld dword [eax + 32]
201 fmul dword [edx + 32]
202 faddp st1,st0
203 fld dword [eax + 36]
204 fmul dword [edx + 36]
205 faddp st1,st0
206 fld dword [eax + 40]
207 fmul dword [edx + 40]
208 faddp st1,st0
209 fld dword [eax + 44]
210 fmul dword [edx + 44]
211 faddp st1,st0
212 fld dword [eax + 48]
213 fmul dword [edx + 48]
214 faddp st1,st0
215 fld dword [eax + 52]
216 fmul dword [edx + 52]
217 faddp st1,st0
218 fld dword [eax + 56]
219 fmul dword [edx + 56]
220 faddp st1,st0
221 fld dword [eax + 60]
222 fmul dword [edx + 60]
223 faddp st1,st0
224 fld dword [eax + 64]
225 fmul dword [edx + 64]
226 faddp st1,st0
227 fld dword [eax + 68]
228 fmul dword [edx + 68]
229 faddp st1,st0
230 fld dword [eax + 72]
231 fmul dword [edx + 72]
232 faddp st1,st0
233 fld dword [eax + 76]
234 fmul dword [edx + 76]
235 faddp st1,st0
236 endproc
237
238
239 proc scalar24_float32_i387
240 %$p arg 4
241 %$q arg 4
242 ;;; alloc
243
244 mov eax,[sp(%$p)]
245 mov edx,[sp(%$q)]
246 fld dword [eax]
247 fmul dword [edx]
248 fld dword [eax + 4]
249 fmul dword [edx + 4]
250 faddp st1,st0
251 fld dword [eax + 8]
252 fmul dword [edx + 8]
253 faddp st1,st0
254 fld dword [eax + 12]
255 fmul dword [edx + 12]
256 faddp st1,st0
257 fld dword [eax + 16]
258 fmul dword [edx + 16]
259 faddp st1,st0
260 fld dword [eax + 20]
261 fmul dword [edx + 20]
262 faddp st1,st0
263 fld dword [eax + 24]
264 fmul dword [edx + 24]
265 faddp st1,st0
266 fld dword [eax + 28]
267 fmul dword [edx + 28]
268 faddp st1,st0
269 fld dword [eax + 32]
270 fmul dword [edx + 32]
271 faddp st1,st0
272 fld dword [eax + 36]
273 fmul dword [edx + 36]
274 faddp st1,st0
275 fld dword [eax + 40]
276 fmul dword [edx + 40]
277 faddp st1,st0
278 fld dword [eax + 44]
279 fmul dword [edx + 44]
280 faddp st1,st0
281 fld dword [eax + 48]
282 fmul dword [edx + 48]
283 faddp st1,st0
284 fld dword [eax + 52]
285 fmul dword [edx + 52]
286 faddp st1,st0
287 fld dword [eax + 56]
288 fmul dword [edx + 56]
289 faddp st1,st0
290 fld dword [eax + 60]
291 fmul dword [edx + 60]
292 faddp st1,st0
293 fld dword [eax + 64]
294 fmul dword [edx + 64]
295 faddp st1,st0
296 fld dword [eax + 68]
297 fmul dword [edx + 68]
298 faddp st1,st0
299 fld dword [eax + 72]
300 fmul dword [edx + 72]
301 faddp st1,st0
302 fld dword [eax + 76]
303 fmul dword [edx + 76]
304 faddp st1,st0
305 fld dword [eax + 80]
306 fmul dword [edx + 80]
307 faddp st1,st0
308 fld dword [eax + 84]
309 fmul dword [edx + 84]
310 faddp st1,st0
311 fld dword [eax + 88]
312 fmul dword [edx + 88]
313 faddp st1,st0
314 fld dword [eax + 92]
315 fmul dword [edx + 92]
316 faddp st1,st0
317 endproc
318
319
320 proc scalar32_float32_i387
321 %$p arg 4
322 %$q arg 4
323 ;;; alloc
324
325 mov eax,[sp(%$p)]
326 mov edx,[sp(%$q)]
327 fld dword [eax]
328 fmul dword [edx]
329 fld dword [eax + 4]
330 fmul dword [edx + 4]
331 faddp st1,st0
332 fld dword [eax + 8]
333 fmul dword [edx + 8]
334 faddp st1,st0
335 fld dword [eax + 12]
336 fmul dword [edx + 12]
337 faddp st1,st0
338 fld dword [eax + 16]
339 fmul dword [edx + 16]
340 faddp st1,st0
341 fld dword [eax + 20]
342 fmul dword [edx + 20]
343 faddp st1,st0
344 fld dword [eax + 24]
345 fmul dword [edx + 24]
346 faddp st1,st0
347 fld dword [eax + 28]
348 fmul dword [edx + 28]
349 faddp st1,st0
350 fld dword [eax + 32]
351 fmul dword [edx + 32]
352 faddp st1,st0
353 fld dword [eax + 36]
354 fmul dword [edx + 36]
355 faddp st1,st0
356 fld dword [eax + 40]
357 fmul dword [edx + 40]
358 faddp st1,st0
359 fld dword [eax + 44]
360 fmul dword [edx + 44]
361 faddp st1,st0
362 fld dword [eax + 48]
363 fmul dword [edx + 48]
364 faddp st1,st0
365 fld dword [eax + 52]
366 fmul dword [edx + 52]
367 faddp st1,st0
368 fld dword [eax + 56]
369 fmul dword [edx + 56]
370 faddp st1,st0
371 fld dword [eax + 60]
372 fmul dword [edx + 60]
373 faddp st1,st0
374 fld dword [eax + 64]
375 fmul dword [edx + 64]
376 faddp st1,st0
377 fld dword [eax + 68]
378 fmul dword [edx + 68]
379 faddp st1,st0
380 fld dword [eax + 72]
381 fmul dword [edx + 72]
382 faddp st1,st0
383 fld dword [eax + 76]
384 fmul dword [edx + 76]
385 faddp st1,st0
386 fld dword [eax + 80]
387 fmul dword [edx + 80]
388 faddp st1,st0
389 fld dword [eax + 84]
390 fmul dword [edx + 84]
391 faddp st1,st0
392 fld dword [eax + 88]
393 fmul dword [edx + 88]
394 faddp st1,st0
395 fld dword [eax + 92]
396 fmul dword [edx + 92]
397 faddp st1,st0
398 fld dword [eax + 96]
399 fmul dword [edx + 96]
400 faddp st1,st0
401 fld dword [eax +100]
402 fmul dword [edx +100]
403 faddp st1,st0
404 fld dword [eax +104]
405 fmul dword [edx +104]
406 faddp st1,st0
407 fld dword [eax +108]
408 fmul dword [edx +108]
409 faddp st1,st0
410 fld dword [eax +112]
411 fmul dword [edx +112]
412 faddp st1,st0
413 fld dword [eax +116]
414 fmul dword [edx +116]
415 faddp st1,st0
416 fld dword [eax +120]
417 fmul dword [edx +120]
418 faddp st1,st0
419 fld dword [eax +124]
420 fmul dword [edx +124]
421 faddp st1,st0
422 endproc
423
424
425 ; float_t scalar4n_float32_i387 (
426 ; const float32_t* const p,
427 ; const float32_t* const q,
428 ; const size_t len );
429
430 proc scalar4n_float32_i387
431 %$p arg 4
432 %$q arg 4
433 %$len arg 4
434 ;;; alloc
435
436 mov eax,[sp(%$p)]
437 mov edx,[sp(%$q)]
438 mov ecx,[sp(%$len)]
439 fld dword [eax]
440 fmul dword [edx]
441 fld dword [eax + 4]
442 fmul dword [edx + 4]
443 faddp st1,st0
444 fld dword [eax + 8]
445 fmul dword [edx + 8]
446 faddp st1,st0
447 fld dword [eax + 12]
448 fmul dword [edx + 12]
449 faddp st1,st0
450 dec ecx
451 jz .ret1
452 add eax,byte 16
453 add edx,byte 16
454 .lbl1
455 fld dword [eax]
456 fmul dword [edx]
457 faddp st1,st0
458 fld dword [eax + 4]
459 fmul dword [edx + 4]
460 faddp st1,st0
461 fld dword [eax + 8]
462 fmul dword [edx + 8]
463 faddp st1,st0
464 fld dword [eax + 12]
465 fmul dword [edx + 12]
466 faddp st1,st0
467 add eax,byte 16
468 add edx,byte 16
469 dec ecx
470 jnz .lbl1
471 .ret1
472 endproc
473
474
475 ; float_t scalar1n_float32_i387 (
476 ; const float32_t* const p,
477 ; const float32_t* const q,
478 ; const size_t len );
479
480 proc scalar1n_float32_i387
481 %$p arg 4
482 %$q arg 4
483 %$len arg 4
484 ;;; alloc
485
486 mov eax,[sp(%$p)]
487 mov edx,[sp(%$q)]
488 mov ecx,[sp(%$len)]
489 fld0
490 shr ecx,1
491 jnc .lbl2
492 fld dword [eax]
493 fmul dword [edx]
494 faddp st1,st0
495 add eax,byte 4
496 add edx,byte 4
497 .lbl2
498 shr ecx,1
499 jnc .lbl3
500 fld dword [eax]
501 fmul dword [edx]
502 faddp st1,st0
503 fld dword [eax + 4]
504 fmul dword [edx + 4]
505 faddp st1,st0
506 add eax,byte 8
507 add edx,byte 8
508 and ecx,ecx
509 .lbl3
510 jz .ret2
511 .lbl4
512 fld dword [eax]
513 fmul dword [edx]
514 faddp st1,st0
515 fld dword [eax + 4]
516 fmul dword [edx + 4]
517 faddp st1,st0
518 fld dword [eax + 8]
519 fmul dword [edx + 8]
520 faddp st1,st0
521 fld dword [eax + 12]
522 fmul dword [edx + 12]
523 faddp st1,st0
524 add eax,byte 16
525 add edx,byte 16
526 dec ecx
527 jnz .lbl4
528 .ret2
529 endproc
530
531
532 proc scalar04_float32_3DNow
533 %$p arg 4
534 %$q arg 4
535 mov eax,[sp(%$p)]
536 mov edx,[sp(%$q)]
537
538 pmov mm0,qword [eax]
539 pmov mm1,qword [eax+8]
540 pfmul mm0,qword [edx]
541 pfmul mm1,qword [edx+8]
542
543 pfadd mm0,mm1
544 pmov qword [sp(%$p)],mm0
545 femms
546 fld dword [sp(%$p)]
547 fadd dword [sp(%$p)+4]
548 endproc
549
550
551 proc scalar08_float32_3DNow
552 %$p arg 4
553 %$q arg 4
554 mov eax,[sp(%$p)]
555 mov edx,[sp(%$q)]
556
557 pmov mm0,qword [eax]
558 pmov mm1,qword [eax+8]
559 pfmul mm0,qword [edx]
560 pfmul mm1,qword [edx+8]
561
562 pmov mm2,qword [eax+16]
563 pmov mm3,qword [eax+24]
564 pfmul mm2,qword [edx+16]
565 pfmul mm3,qword [edx+24]
566 pfadd mm0,mm2
567 pfadd mm1,mm3
568
569 pfadd mm0,mm1
570 pmov qword [sp(%$p)],mm0
571 femms
572 fld dword [sp(%$p)]
573 fadd dword [sp(%$p)+4]
574 endproc
575
576
577 proc scalar12_float32_3DNow
578 %$p arg 4
579 %$q arg 4
580 mov eax,[sp(%$p)]
581 mov edx,[sp(%$q)]
582
583 pmov mm0,qword [eax]
584 pmov mm1,qword [eax+8]
585 pfmul mm0,qword [edx]
586 pfmul mm1,qword [edx+8]
587
588 pmov mm2,qword [eax+16]
589 pmov mm3,qword [eax+24]
590 pfmul mm2,qword [edx+16]
591 pfmul mm3,qword [edx+24]
592 pfadd mm0,mm2
593 pfadd mm1,mm3
594
595 pmov mm2,qword [eax+32]
596 pmov mm3,qword [eax+40]
597 pfmul mm2,qword [edx+32]
598 pfmul mm3,qword [edx+40]
599 pfadd mm0,mm2
600 pfadd mm1,mm3
601
602 pfadd mm0,mm1
603 pmov qword [sp(%$p)],mm0
604 femms
605 fld dword [sp(%$p)]
606 fadd dword [sp(%$p)+4]
607 endproc
608
609
610 proc scalar16_float32_3DNow
611 %$p arg 4
612 %$q arg 4
613 mov eax,[sp(%$p)]
614 mov edx,[sp(%$q)]
615
616 pmov mm0,qword [eax]
617 pmov mm1,qword [eax+8]
618 pfmul mm0,qword [edx]
619 pfmul mm1,qword [edx+8]
620
621 pmov mm2,qword [eax+16]
622 pmov mm3,qword [eax+24]
623 pfmul mm2,qword [edx+16]
624 pfmul mm3,qword [edx+24]
625 pfadd mm0,mm2
626 pfadd mm1,mm3
627
628 pmov mm2,qword [eax+32]
629 pmov mm3,qword [eax+40]
630 pfmul mm2,qword [edx+32]
631 pfmul mm3,qword [edx+40]
632 pfadd mm0,mm2
633 pfadd mm1,mm3
634
635 pmov mm2,qword [eax+48]
636 pmov mm3,qword [eax+56]
637 pfmul mm2,qword [edx+48]
638 pfmul mm3,qword [edx+56]
639 pfadd mm0,mm2
640 pfadd mm1,mm3
641
642 pfadd mm0,mm1
643 pmov qword [sp(%$p)],mm0
644 femms
645 fld dword [sp(%$p)]
646 fadd dword [sp(%$p)+4]
647 endproc
648
649
650 proc scalar20_float32_3DNow
651 %$p arg 4
652 %$q arg 4
653 mov eax,[sp(%$p)]
654 mov edx,[sp(%$q)]
655
656 pmov mm0,qword [eax]
657 pmov mm1,qword [eax+8]
658 pfmul mm0,qword [edx]
659 pfmul mm1,qword [edx+8]
660
661 pmov mm2,qword [eax+16]
662 pmov mm3,qword [eax+24]
663 pfmul mm2,qword [edx+16]
664 pfmul mm3,qword [edx+24]
665 pfadd mm0,mm2
666 pfadd mm1,mm3
667
668 pmov mm2,qword [eax+32]
669 pmov mm3,qword [eax+40]
670 pfmul mm2,qword [edx+32]
671 pfmul mm3,qword [edx+40]
672 pfadd mm0,mm2
673 pfadd mm1,mm3
674
675 pmov mm2,qword [eax+48]
676 pmov mm3,qword [eax+56]
677 pfmul mm2,qword [edx+48]
678 pfmul mm3,qword [edx+56]
679 pfadd mm0,mm2
680 pfadd mm1,mm3
681
682 pmov mm2,qword [eax+64]
683 pmov mm3,qword [eax+72]
684 pfmul mm2,qword [edx+64]
685 pfmul mm3,qword [edx+72]
686 pfadd mm0,mm2
687 pfadd mm1,mm3
688
689 pfadd mm0,mm1
690 pmov qword [sp(%$p)],mm0
691 femms
692 fld dword [sp(%$p)]
693 fadd dword [sp(%$p)+4]
694 endproc
695
696
697 proc scalar24_float32_3DNow
698 %$p arg 4
699 %$q arg 4
700 mov eax,[sp(%$p)]
701 mov edx,[sp(%$q)]
702
703 pmov mm0,qword [eax]
704 pmov mm1,qword [eax+8]
705 pfmul mm0,qword [edx]
706 pfmul mm1,qword [edx+8]
707
708 pmov mm2,qword [eax+16]
709 pmov mm3,qword [eax+24]
710 pfmul mm2,qword [edx+16]
711 pfmul mm3,qword [edx+24]
712 pfadd mm0,mm2
713 pfadd mm1,mm3
714
715 pmov mm2,qword [eax+32]
716 pmov mm3,qword [eax+40]
717 pfmul mm2,qword [edx+32]
718 pfmul mm3,qword [edx+40]
719 pfadd mm0,mm2
720 pfadd mm1,mm3
721
722 pmov mm2,qword [eax+48]
723 pmov mm3,qword [eax+56]
724 pfmul mm2,qword [edx+48]
725 pfmul mm3,qword [edx+56]
726 pfadd mm0,mm2
727 pfadd mm1,mm3
728
729 pmov mm2,qword [eax+64]
730 pmov mm3,qword [eax+72]
731 pfmul mm2,qword [edx+64]
732 pfmul mm3,qword [edx+72]
733 pfadd mm0,mm2
734 pfadd mm1,mm3
735
736 pmov mm2,qword [eax+80]
737 pmov mm3,qword [eax+88]
738 pfmul mm2,qword [edx+80]
739 pfmul mm3,qword [edx+88]
740 pfadd mm0,mm2
741 pfadd mm1,mm3
742
743 pfadd mm0,mm1
744 pmov qword [sp(%$p)],mm0
745 femms
746 fld dword [sp(%$p)]
747 fadd dword [sp(%$p)+4]
748 endproc
749
750 proc scalar32_float32_3DNow
751 %$p arg 4
752 %$q arg 4
753 mov eax,[sp(%$p)]
754 mov edx,[sp(%$q)]
755
756 pmov mm0,qword [eax]
757 pmov mm1,qword [eax+8]
758 pfmul mm0,qword [edx]
759 pfmul mm1,qword [edx+8]
760
761 pmov mm2,qword [eax+16]
762 pmov mm3,qword [eax+24]
763 pfmul mm2,qword [edx+16]
764 pfmul mm3,qword [edx+24]
765 pfadd mm0,mm2
766 pfadd mm1,mm3
767
768 pmov mm2,qword [eax+32]
769 pmov mm3,qword [eax+40]
770 pfmul mm2,qword [edx+32]
771 pfmul mm3,qword [edx+40]
772 pfadd mm0,mm2
773 pfadd mm1,mm3
774
775 pmov mm2,qword [eax+48]
776 pmov mm3,qword [eax+56]
777 pfmul mm2,qword [edx+48]
778 pfmul mm3,qword [edx+56]
779 pfadd mm0,mm2
780 pfadd mm1,mm3
781
782 pmov mm2,qword [eax+64]
783 pmov mm3,qword [eax+72]
784 pfmul mm2,qword [edx+64]
785 pfmul mm3,qword [edx+72]
786 pfadd mm0,mm2
787 pfadd mm1,mm3
788
789 pmov mm2,qword [eax+80]
790 pmov mm3,qword [eax+88]
791 pfmul mm2,qword [edx+80]
792 pfmul mm3,qword [edx+88]
793 pfadd mm0,mm2
794 pfadd mm1,mm3
795
796 pmov mm2,qword [eax+96]
797 pmov mm3,qword [eax+104]
798 pfmul mm2,qword [edx+96]
799 pfmul mm3,qword [edx+104]
800 pfadd mm0,mm2
801 pfadd mm1,mm3
802
803 pmov mm2,qword [eax+112]
804 pmov mm3,qword [eax+120]
805 pfmul mm2,qword [edx+112]
806 pfmul mm3,qword [edx+120]
807 pfadd mm0,mm2
808 pfadd mm1,mm3
809
810 pfadd mm0,mm1
811 pmov qword [sp(%$p)],mm0
812 femms
813 fld dword [sp(%$p)]
814 fadd dword [sp(%$p)+4]
815 endproc
816
817
818 proc scalar4n_float32_3DNow
819 %$p arg 4
820 %$q arg 4
821 %$len arg 4
822
823 mov eax,[sp(%$p)]
824 mov edx,[sp(%$q)]
825 mov ecx,[sp(%$len)]
826
827 pmov mm0,qword [eax]
828 pmov mm1,qword [eax+8]
829 pfmul mm0,qword [edx]
830 pfmul mm1,qword [edx+8]
831 dec ecx
832 jz .ret4
833
834 add eax,byte 16
835 add edx,byte 16
836 .lbl4:
837 pmov mm2,qword [eax]
838 pmov mm3,qword [eax+8]
839 pfmul mm2,qword [edx]
840 pfmul mm3,qword [edx+8]
841 add eax,byte 16
842 add edx,byte 16
843 pfadd mm0,mm2
844 pfadd mm1,mm3
845 dec ecx
846 jnz .lbl4
847
848 .ret4: pfadd mm0,mm1
849 pmov qword [sp(%$p)],mm0
850 femms
851 fld dword [sp(%$p)]
852 fadd dword [sp(%$p)+4]
853 endproc
854
855
856 proc scalar1n_float32_3DNow
857 jmp scalar24_float32_i387
858 endproc
859
860
861 proc scalar04_float32_SIMD
862 jmp scalar04_float32_i387
863 endproc
864
865
866 proc scalar08_float32_SIMD
867 %$p arg 4
868 %$q arg 4
869 mov eax,[sp(%$p)]
870 mov edx,[sp(%$q)]
871
872 movups xmm0, [eax]
873 movups xmm1, [eax+16]
874 mulps xmm0, [edx]
875 mulps xmm1, [edx+16]
876
877 addps xmm0,xmm1
878 sub esp,16
879 movups [esp],xmm0
880 fld dword [esp+ 0]
881 fadd dword [esp+ 4]
882 fadd dword [esp+ 8]
883 fadd dword [esp+12]
884 add esp,16
885 endproc
886
887
888 proc scalar12_float32_SIMD
889 jmp scalar12_float32_i387
890 endproc
891
892
893 proc scalar16_float32_SIMD
894 %$p arg 4
895 %$q arg 4
896 mov eax,[sp(%$p)]
897 mov edx,[sp(%$q)]
898
899 movups xmm0, [eax]
900 movups xmm1, [eax+16]
901 mulps xmm0, [edx]
902 mulps xmm1, [edx+16]
903
904 movups xmm2, [eax+32]
905 movups xmm3, [eax+48]
906 mulps xmm2, [edx+32]
907 mulps xmm3, [edx+48]
908 addps xmm0,xmm2
909 addps xmm1,xmm3
910
911 addps xmm0,xmm1
912 sub esp,16
913 movups [esp],xmm0
914 fld dword [esp+ 0]
915 fadd dword [esp+ 4]
916 fadd dword [esp+ 8]
917 fadd dword [esp+12]
918 add esp,16
919 endproc
920
921
922 proc scalar20_float32_SIMD
923 jmp scalar20_float32_i387
924 endproc
925
926
927 proc scalar24_float32_SIMD
928 %$p arg 4
929 %$q arg 4
930 mov eax,[sp(%$p)]
931 mov edx,[sp(%$q)]
932
933 movups xmm0, [eax]
934 movups xmm1, [eax+16]
935 mulps xmm0, [edx]
936 mulps xmm1, [edx+16]
937
938 movups xmm2, [eax+32]
939 movups xmm3, [eax+48]
940 mulps xmm2, [edx+32]
941 mulps xmm3, [edx+48]
942 addps xmm0,xmm2
943 addps xmm1,xmm3
944
945 movups xmm2, [eax+64]
946 movups xmm3, [eax+80]
947 mulps xmm2, [edx+64]
948 mulps xmm3, [edx+80]
949 addps xmm0,xmm2
950 addps xmm1,xmm3
951
952 addps xmm0,xmm1
953 sub esp,16
954 movups [esp],xmm0
955 fld dword [esp+ 0]
956 fadd dword [esp+ 4]
957 fadd dword [esp+ 8]
958 fadd dword [esp+12]
959 add esp,16
960 endproc
961
962
963 proc scalar32_float32_SIMD
964 %$p arg 4
965 %$q arg 4
966 mov eax,[sp(%$p)]
967 mov edx,[sp(%$q)]
968
969 movups xmm0, [eax]
970 movups xmm1, [eax+16]
971 mulps xmm0, [edx]
972 mulps xmm1, [edx+16]
973
974 movups xmm2, [eax+32]
975 movups xmm3, [eax+48]
976 mulps xmm2, [edx+32]
977 mulps xmm3, [edx+48]
978 addps xmm0,xmm2
979 addps xmm1,xmm3
980
981 movups xmm2, [eax+64]
982 movups xmm3, [eax+80]
983 mulps xmm2, [edx+64]
984 mulps xmm3, [edx+80]
985 addps xmm0,xmm2
986 addps xmm1,xmm3
987
988 movups xmm2, [eax+96]
989 movups xmm3, [eax+112]
990 mulps xmm2, [edx+96]
991 mulps xmm3, [edx+112]
992 addps xmm0,xmm2
993 addps xmm1,xmm3
994
995 addps xmm0,xmm1
996
997 ;sub esp,16
998 ;movups [esp],xmm0
999 ;fld dword [esp+ 0]
1000 ;fadd dword [esp+ 4]
1001 ;fadd dword [esp+ 8]
1002 ;fadd dword [esp+12]
1003 ;add esp,16
1004
1005 movhlps xmm1,xmm0
1006 addps xmm0,xmm1
1007 movlps [sp(%$p)],xmm0
1008 fld dword [sp(%$p)]
1009 fadd dword [sp(%$p)+4]
1010 endproc
1011
1012
1013 proc scalar4n_float32_SIMD
1014 jmp scalar4n_float32_i387
1015 endproc
1016
1017
1018 proc scalar1n_float32_SIMD
1019 jmp scalar1n_float32_i387
1020 endproc
1021
1022 ; end of scalar.nas