Lines Matching refs:QWP
71 $Asse2=&QWP(0,"esp");
72 $Bsse2=&QWP(8,"esp");
73 $Csse2=&QWP(16,"esp");
74 $Dsse2=&QWP(24,"esp");
75 $Esse2=&QWP(32,"esp");
76 $Fsse2=&QWP(40,"esp");
77 $Gsse2=&QWP(48,"esp");
78 $Hsse2=&QWP(56,"esp");
98 &movq (&QWP(8*9,"esp"),"mm7") # save X[i]
112 &paddq ("mm7",QWP(0,$K512)); # X[i]+=K512[i]
137 &movq ("mm7",&QWP(8*(9+16-1),"esp")) if ($phase!=0); # pre-fetch
313 &movq ($A,&QWP(0,"esi"));
315 &movq ("mm1",&QWP(8,"esi"));
317 &movq ($BxC,&QWP(16,"esi"));
319 &movq ("mm3",&QWP(24,"esi"));
320 &movq ($E,&QWP(32,"esi"));
321 &movq ("mm5",&QWP(40,"esi"));
322 &movq ("mm6",&QWP(48,"esi"));
323 &movq ("mm7",&QWP(56,"esi"));
376 #&movq ("mm7",&QWP(8*(9+16-1),"esp")); # prefetched in BODY_00_15
377 &movq ("mm5",&QWP(8*(9+16-14),"esp"));
397 &paddq ("mm7",&QWP(8*(9+16),"esp"));
400 &paddq ("mm7",&QWP(8*(9+16-9),"esp"));
408 #&movq (&QWP(8*9,"esp"),"mm7"); # moved to BODY_00_15
426 &paddq ($A,&QWP(0,"esi"));
427 &paddq ("mm1",&QWP(8,"esi"));
428 &paddq ($BxC,&QWP(16,"esi"));
429 &paddq ("mm3",&QWP(24,"esi"));
430 &paddq ($E,&QWP(32,"esi"));
431 &paddq ("mm5",&QWP(40,"esi"));
432 &paddq ("mm6",&QWP(48,"esi"));
433 &paddq ("mm7",&QWP(56,"esi"));
436 &movq (&QWP(0,"esi"),$A);
437 &movq (&QWP(8,"esi"),"mm1");
438 &movq (&QWP(16,"esi"),$BxC);
439 &movq (&QWP(24,"esi"),"mm3");
440 &movq (&QWP(32,"esi"),$E);
441 &movq (&QWP(40,"esi"),"mm5");
442 &movq (&QWP(48,"esi"),"mm6");
443 &movq (&QWP(56,"esi"),"mm7");
471 &movdqa (@X[1],&QWP(80*8,$K512)); # byte swap mask
472 &movdqu (@X[0],&QWP(0,"edi"));
475 &movdqa (&QWP(16*(($j-1)%4),$frame),@X[3]) if ($j>4); # off-load
476 &movdqa (@X[3],&QWP(16*($j%8),$K512));
478 &movdqu (@X[1],&QWP(16*($j+1),"edi")) if ($j<7); # next input
479 &movdqa (@X[1],&QWP(16*(($j+1)%4),$frame)) if ($j==7);# restore @X[0]
482 &movdqa (&QWP(16*($j%8)-128,$frame),@X[3]); # xfer X[i]+K[i]
490 &movdqa (@X[2],&QWP(16*(($j+1)%4),$frame)); # pre-restore @X[1]
491 &movdqa (&QWP(16*(($j-1)%4),$frame),@X[3]); # off-load @X[3]
515 '&movq ("mm7",&QWP(((-8*$i)%128)-128,$frame))',# X[i]+K[i]
518 '&movq (&QWP(8*($i+4)%64,"esp"),$E)', # modulo-scheduled save e
528 '&movq (&QWP(8*$i%64,"esp"),$A)', # modulo-scheduled save a
532 '&paddq("mm7",&QWP(8*($i+7)%64,"esp"))', # X[i]+=h
537 '&movq ($E,&QWP(8*($i+3)%64,"esp"))', # e = load d, e in next round
545 '&movq ("mm1",&QWP(8*($i+1)%64,"esp"))', # load b
558 '&movq ("mm5",&QWP(8*($i+5-1)%64,"esp"))', # pre-load f
560 '&movq ("mm6",&QWP(8*($i+6-1)%64,"esp"))', # pre-load g
575 &movdqa (&QWP(16*($j%4),$frame),@X[4]); # off-load @X[4]
603 &movdqa ($t2,&QWP(16*(($j+2)%4),$frame));# pre-restore @X[1]
605 &movdqa ($t0,&QWP(16*($j%8),$K512));
614 &movdqa (&QWP(16*($j%8)-128,$frame),$t0);# xfer X[i]+K[i]
622 &movdqa (@X[1],&QWP(0,$K512)); # byte swap mask
624 &movdqu (@X[0],&QWP(0,"ebx"));
630 &movdqa (&QWP(16*(($j-1)%4),$frame),@X[3]) if ($j>4); # off-load
631 &movdqa (@X[3],&QWP(16*($j%8),$K512));
633 &movdqu (@X[1],&QWP(16*($j+1),"ebx")) if ($j<7); # next input
634 &movdqa (@X[1],&QWP(16*(($j+1)%4),$frame)) if ($j==7);# restore @X[0]
638 &movdqa (&QWP(16*($j%8)-128,$frame),@X[3]);# xfer X[i]+K[i]
654 &paddq ($A,&QWP(0,"esi"));
655 &paddq ("mm1",&QWP(8,"esi"));
656 &paddq ($BxC,&QWP(16,"esi"));
657 &paddq ("mm3",&QWP(24,"esi"));
658 &paddq ($E,&QWP(32,"esi"));
659 &paddq ("mm5",&QWP(40,"esi"));
660 &paddq ("mm6",&QWP(48,"esi"));
661 &paddq ("mm7",&QWP(56,"esi"));
663 &movq (&QWP(0,"esi"),$A);
664 &movq (&QWP(8,"esi"),"mm1");
665 &movq (&QWP(16,"esi"),$BxC);
666 &movq (&QWP(24,"esi"),"mm3");
667 &movq (&QWP(32,"esi"),$E);
668 &movq (&QWP(40,"esi"),"mm5");
669 &movq (&QWP(48,"esi"),"mm6");
670 &movq (&QWP(56,"esi"),"mm7");