Lines Matching full:i
33 fHIDE(int i;) \
34 fVFOREACH(WIDTH, i) {\
214 DSTM(0,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VvV.SRCTYPE[i],shamt) >> shamt)); \
215 DSTM(1,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VuV.SRCTYPE[i],shamt) >> shamt)))
218 …ESCR, VdV.DEST[i] = fVAVGS( WIDTH, VuV.SRC[i], VvV.SR…
219 …Round"DESCR, VdV.DEST[i] = fVAVGSRND( WIDTH, VuV.SRC[i], VvV.SR…
220 …verage "DESCR, VdV.DEST[i] = fVNAVGS( WIDTH, VuV.SRC[i], VvV.SR…
223 …"DESCR, VdV.DEST[i] = fVAVGU( WIDTH, VuV.SRC[i], VvV.SRC[i…
224 …% Round"DESCR, VdV.DEST[i] = fVAVGURND(WIDTH, VuV.SRC[i], VvV.SRC[i…
340 fHIDE(int i;) \
341 for(i = 0; i < fVBYTES(); i++) {\
342 VdV.ub[i] = (i+SHIFT>=fVBYTES()) ? VuV.ub[i+SHIFT-fVBYTES()] : VvV.ub[i+SHIFT];\
384 …)", "Unpack byte with zero-extend", fVARRAY_ELEMENT_ACCESS(VddV, uh, i) = fZE8_16( VuV.ub[i]))
385 …, "Unpack bytes with sign-extend", fVARRAY_ELEMENT_ACCESS(VddV, h, i) = fSE8_16( VuV.b[i] ))
386 …)", "Unpack halves with zero-extend", fVARRAY_ELEMENT_ACCESS(VddV, uw, i) = fZE16_32(VuV.uh[i]))
387 …, "Unpack halves with sign-extend", fVARRAY_ELEMENT_ACCESS(VddV, w, i) = fSE16_32(VuV.h[i] ))
389 …, "Unpack byte to odd bytes ", fVARRAY_ELEMENT_ACCESS(VxxV, uh, i) |= fZE8_16( VuV.ub[i])<<8)
390 … "Unpack halves to odd halves", fVARRAY_ELEMENT_ACCESS(VxxV, uw, i) |= fZE16_32(VuV.uh[i])<<16)
399 VdV.ub[i] = fGETUBYTE(0, VvV.uh[i]);
400 VdV.ub[i+fVELEM(16)] = fGETUBYTE(0, VuV.uh[i]))
404 VdV.uh[i] = fGETUHALF(0, VvV.uw[i]);
405 VdV.uh[i+fVELEM(32)] = fGETUHALF(0, VuV.uw[i]))
409 VdV.ub[i] = fGETUBYTE(1, VvV.uh[i]);
410 VdV.ub[i+fVELEM(16)] = fGETUBYTE(1, VuV.uh[i]))
414 VdV.uh[i] = fGETUHALF(1, VvV.uw[i]);
415 VdV.uh[i+fVELEM(32)] = fGETUHALF(1, VuV.uw[i]))
421 VdV.ub[i] = fVSATUB(VvV.h[i]);
422 VdV.ub[i+fVELEM(16)] = fVSATUB(VuV.h[i]))
427 VdV.b[i] = fVSATB(VvV.h[i]);
428 VdV.b[i+fVELEM(16)] = fVSATB(VuV.h[i]))
433 VdV.uh[i] = fVSATUH(VvV.w[i]);
434 VdV.uh[i+fVELEM(32)] = fVSATUH(VuV.w[i]))
438 VdV.h[i] = fVSATH(VvV.w[i]);
439 VdV.h[i+fVELEM(32)] = fVSATH(VuV.w[i]))
451 VddV.v[0].uh[i] = fZE8_16(fGETUBYTE(0, VuV.uh[i]));
452 VddV.v[1].uh[i] = fZE8_16(fGETUBYTE(1, VuV.uh[i])))
456 VddV.v[0].h[i] = fSE8_16(fGETBYTE(0, VuV.h[i]));
457 VddV.v[1].h[i] = fSE8_16(fGETBYTE(1, VuV.h[i])))
461 VddV.v[0].uw[i] = fZE16_32(fGETUHALF(0, VuV.uw[i]));
462 VddV.v[1].uw[i] = fZE16_32(fGETUHALF(1, VuV.uw[i])))
466 VddV.v[0].w[i] = fSE16_32(fGETHALF(0, VuV.w[i]));
467 VddV.v[1].w[i] = fSE16_32(fGETHALF(1, VuV.w[i])))
487 VdV.h[i] = fMPY8US( fGETUBYTE(0, VuV.uh[i]), fGETBYTE((2*i) % 4, RtV));
488 VdV.h[i] += fMPY8US( fGETUBYTE(1, VuV.uh[i]), fGETBYTE((2*i+1)%4, RtV)))
492 VxV.h[i] += fMPY8US( fGETUBYTE(0, VuV.uh[i]), fGETBYTE((2*i) % 4, RtV));
493 VxV.h[i] += fMPY8US( fGETUBYTE(1, VuV.uh[i]), fGETBYTE((2*i+1)%4, RtV)))
499 VddV.v[0].h[i] = fMPY8US(fGETUBYTE(0, VuuV.v[0].uh[i]),fGETBYTE((2*i) % 4, RtV));
500 VddV.v[0].h[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]),fGETBYTE((2*i+1)%4, RtV));
502 VddV.v[1].h[i] = fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]),fGETBYTE((2*i) % 4, RtV));
503 VddV.v[1].h[i] += fMPY8US(fGETUBYTE(0, VuuV.v[1].uh[i]),fGETBYTE((2*i+1)%4, RtV)))
507 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(0, VuuV.v[0].uh[i]),fGETBYTE((2*i) % 4, RtV));
508 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]),fGETBYTE((2*i+1)%4, RtV));
510 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]),fGETBYTE((2*i) % 4, RtV));
511 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(0, VuuV.v[1].uh[i]),fGETBYTE((2*i+1)%4, RtV)))
520 VdV.w[i] = fMPY16SS(fGETHALF(0, VuV.w[i]),fGETBYTE((2*i+0)%4, RtV));
521 VdV.w[i] += fMPY16SS(fGETHALF(1, VuV.w[i]),fGETBYTE((2*i+1)%4, RtV)))
525 VxV.w[i] += fMPY16SS(fGETHALF(0, VuV.w[i]),fGETBYTE((2*i+0)%4, RtV));
526 VxV.w[i] += fMPY16SS(fGETHALF(1, VuV.w[i]),fGETBYTE((2*i+1)%4, RtV)))
532 VddV.v[0].w[i] = fMPY16SS(fGETHALF(0, VuuV.v[0].w[i]),fGETBYTE((2*i+0)%4, RtV));
533 VddV.v[0].w[i] += fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]),fGETBYTE((2*i+1)%4, RtV));
535 VddV.v[1].w[i] = fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]),fGETBYTE((2*i+0)%4, RtV));
536 VddV.v[1].w[i] += fMPY16SS(fGETHALF(0, VuuV.v[1].w[i]),fGETBYTE((2*i+1)%4, RtV)))
541 VxxV.v[0].w[i] += fMPY16SS(fGETHALF(0, VuuV.v[0].w[i]),fGETBYTE((2*i+0)%4, RtV));
542 VxxV.v[0].w[i] += fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]),fGETBYTE((2*i+1)%4, RtV));
544 VxxV.v[1].w[i] += fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]),fGETBYTE((2*i+0)%4, RtV));
545 VxxV.v[1].w[i] += fMPY16SS(fGETHALF(0, VuuV.v[1].w[i]),fGETBYTE((2*i+1)%4, RtV)))
558 accum = fMPY16SS(fGETHALF(0,VuV.w[i]),fGETHALF(0, VvV.w[i]));
559 accum += fMPY16SS(fGETHALF(1,VuV.w[i]),fGETHALF(1, VvV.w[i]));
560 VdV.w[i] = fVSATW(accum))
565 accum = fMPY16SS(fGETHALF(0,VuV.w[i]),fGETHALF(0, VvV.w[i]));
566 accum += fMPY16SS(fGETHALF(1,VuV.w[i]),fGETHALF(1, VvV.w[i]));
567 VxV.w[i] = fVSATW(VxV.w[i]+accum))
575 accum = fMPY16SS(fGETHALF(0, VuV.w[i]),fGETHALF(0, RtV));
576 accum += fMPY16SS(fGETHALF(1, VuV.w[i]),fGETHALF(1, RtV));
577 VdV.w[i] = fVSATW(accum))
581 fHIDE(size8s_t) accum = VxV.w[i];
582 accum += fMPY16SS(fGETHALF(0, VuV.w[i]),fGETHALF(0, RtV));
583 accum += fMPY16SS(fGETHALF(1, VuV.w[i]),fGETHALF(1, RtV));
584 VxV.w[i] = fVSATW(accum))
592 accum = fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]),fGETHALF(0,RtV));
593 accum += fMPY16SS(fGETHALF(0,VuuV.v[1].w[i]),fGETHALF(1,RtV));
594 VdV.w[i] = fVSATW(accum))
598 fHIDE(size8s_t) accum = VxV.w[i];
599 accum += fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]),fGETHALF(0,RtV));
600 accum += fMPY16SS(fGETHALF(0,VuuV.v[1].w[i]),fGETHALF(1,RtV));
601 VxV.w[i] = fVSATW(accum))
613 accum = fMPY16SU(fGETHALF(0, VuV.w[i]),fGETUHALF(0, RtV));
614 accum += fMPY16SU(fGETHALF(1, VuV.w[i]),fGETUHALF(1, RtV));
615 VdV.w[i] = fVSATW(accum))
619 fHIDE(size8s_t) accum=VxV.w[i];
620 accum += fMPY16SU(fGETHALF(0, VuV.w[i]),fGETUHALF(0, RtV));
621 accum += fMPY16SU(fGETHALF(1, VuV.w[i]),fGETUHALF(1, RtV));
622 VxV.w[i] = fVSATW(accum))
629 accum = fMPY16SU(fGETHALF(1,VuuV.v[0].w[i]),fGETUHALF(0,RtV));
630 accum += fMPY16SU(fGETHALF(0,VuuV.v[1].w[i]),fGETUHALF(1,RtV));
631 VdV.w[i] = fVSATW(accum))
635 fHIDE(size8s_t) accum=VxV.w[i];
636 accum += fMPY16SU(fGETHALF(1, VuuV.v[0].w[i]),fGETUHALF(0,RtV));
637 accum += fMPY16SU(fGETHALF(0, VuuV.v[1].w[i]),fGETUHALF(1,RtV));
638 VxV.w[i] = fVSATW(accum))
648 VddV.v[0].h[i] = fMPY8SS(fGETBYTE(0,VuuV.v[0].h[i]), fGETBYTE((2*i )%4, RtV));
649 VddV.v[0].h[i] += fMPY8SS(fGETBYTE(1,VuuV.v[0].h[i]), fGETBYTE((2*i+1)%4, RtV));
650 VddV.v[0].h[i] += fGETBYTE(0,VuuV.v[1].h[i]);
652 VddV.v[1].h[i] = fMPY8SS(fGETBYTE(1,VuuV.v[0].h[i]), fGETBYTE((2*i )%4, RtV));
653 VddV.v[1].h[i] += fMPY8SS(fGETBYTE(0,VuuV.v[1].h[i]), fGETBYTE((2*i+1)%4, RtV));
654 VddV.v[1].h[i] += fGETBYTE(1,VuuV.v[1].h[i]))
659 VxxV.v[0].h[i] += fMPY8SS(fGETBYTE(0,VuuV.v[0].h[i]), fGETBYTE((2*i )%4, RtV));
660 VxxV.v[0].h[i] += fMPY8SS(fGETBYTE(1,VuuV.v[0].h[i]), fGETBYTE((2*i+1)%4, RtV));
661 VxxV.v[0].h[i] += fGETBYTE(0,VuuV.v[1].h[i]);
663 VxxV.v[1].h[i] += fMPY8SS(fGETBYTE(1,VuuV.v[0].h[i]), fGETBYTE((2*i )%4, RtV));
664 VxxV.v[1].h[i] += fMPY8SS(fGETBYTE(0,VuuV.v[1].h[i]), fGETBYTE((2*i+1)%4, RtV));
665 VxxV.v[1].h[i] += fGETBYTE(1,VuuV.v[1].h[i]))
671 VddV.v[0].h[i] = fMPY8US(fGETUBYTE(0,VuuV.v[0].uh[i]), fGETBYTE((2*i )%4, RtV));
672 VddV.v[0].h[i] += fMPY8US(fGETUBYTE(1,VuuV.v[0].uh[i]), fGETBYTE((2*i+1)%4, RtV));
673 VddV.v[0].h[i] += fGETUBYTE(0,VuuV.v[1].uh[i]);
675 VddV.v[1].h[i] = fMPY8US(fGETUBYTE(1,VuuV.v[0].uh[i]), fGETBYTE((2*i )%4, RtV));
676 VddV.v[1].h[i] += fMPY8US(fGETUBYTE(0,VuuV.v[1].uh[i]), fGETBYTE((2*i+1)%4, RtV));
677 VddV.v[1].h[i] += fGETUBYTE(1,VuuV.v[1].uh[i]))
681 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(0,VuuV.v[0].uh[i]), fGETBYTE((2*i )%4, RtV));
682 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(1,VuuV.v[0].uh[i]), fGETBYTE((2*i+1)%4, RtV));
683 VxxV.v[0].h[i] += fGETUBYTE(0,VuuV.v[1].uh[i]);
685 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(1,VuuV.v[0].uh[i]), fGETBYTE((2*i )%4, RtV));
686 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(0,VuuV.v[1].uh[i]), fGETBYTE((2*i+1)%4, RtV));
687 VxxV.v[1].h[i] += fGETUBYTE(1,VuuV.v[1].uh[i]))
692 VddV.v[0].w[i] = fMPY16SS(fGETHALF(0,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+0)%4, RtV)));
693 VddV.v[0].w[i]+= fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+1)%4, RtV)));
694 VddV.v[0].w[i]+= fGETHALF(0,VuuV.v[1].w[i]);
696 VddV.v[1].w[i] = fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+0)%4, RtV)));
697 VddV.v[1].w[i]+= fMPY16SS(fGETHALF(0,VuuV.v[1].w[i]), fSE8_16(fGETBYTE((2*i+1)%4, RtV)));
698 VddV.v[1].w[i]+= fGETHALF(1,VuuV.v[1].w[i]))
702 VxxV.v[0].w[i]+= fMPY16SS(fGETHALF(0,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+0)%4, RtV)));
703 VxxV.v[0].w[i]+= fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+1)%4, RtV)));
704 VxxV.v[0].w[i]+= fGETHALF(0,VuuV.v[1].w[i]);
706 VxxV.v[1].w[i]+= fMPY16SS(fGETHALF(1,VuuV.v[0].w[i]), fSE8_16(fGETBYTE((2*i+0)%4, RtV)));
707 VxxV.v[1].w[i]+= fMPY16SS(fGETHALF(0,VuuV.v[1].w[i]), fSE8_16(fGETBYTE((2*i+1)%4, RtV)));
708 VxxV.v[1].w[i]+= fGETHALF(1,VuuV.v[1].w[i]))
719 VdV.uw[i] = fMPY8UU(fGETUBYTE(0,VuV.uw[i]), fGETUBYTE(0,RtV));
720 VdV.uw[i] += fMPY8UU(fGETUBYTE(1,VuV.uw[i]), fGETUBYTE(1,RtV));
721 VdV.uw[i] += fMPY8UU(fGETUBYTE(2,VuV.uw[i]), fGETUBYTE(2,RtV));
722 VdV.uw[i] += fMPY8UU(fGETUBYTE(3,VuV.uw[i]), fGETUBYTE(3,RtV)))
726 VxV.uw[i] += fMPY8UU(fGETUBYTE(0,VuV.uw[i]), fGETUBYTE(0,RtV));
727 VxV.uw[i] += fMPY8UU(fGETUBYTE(1,VuV.uw[i]), fGETUBYTE(1,RtV));
728 VxV.uw[i] += fMPY8UU(fGETUBYTE(2,VuV.uw[i]), fGETUBYTE(2,RtV));
729 VxV.uw[i] += fMPY8UU(fGETUBYTE(3,VuV.uw[i]), fGETUBYTE(3,RtV)))
734 VdV.uw[i] = fMPY8UU(fGETUBYTE(0,VuV.uw[i]), fGETUBYTE(0,VvV.uw[i]));
735 VdV.uw[i] += fMPY8UU(fGETUBYTE(1,VuV.uw[i]), fGETUBYTE(1,VvV.uw[i]));
736 VdV.uw[i] += fMPY8UU(fGETUBYTE(2,VuV.uw[i]), fGETUBYTE(2,VvV.uw[i]));
737 VdV.uw[i] += fMPY8UU(fGETUBYTE(3,VuV.uw[i]), fGETUBYTE(3,VvV.uw[i])))
741 VxV.uw[i] += fMPY8UU(fGETUBYTE(0,VuV.uw[i]), fGETUBYTE(0,VvV.uw[i]));
742 VxV.uw[i] += fMPY8UU(fGETUBYTE(1,VuV.uw[i]), fGETUBYTE(1,VvV.uw[i]));
743 VxV.uw[i] += fMPY8UU(fGETUBYTE(2,VuV.uw[i]), fGETUBYTE(2,VvV.uw[i]));
744 VxV.uw[i] += fMPY8UU(fGETUBYTE(3,VuV.uw[i]), fGETUBYTE(3,VvV.uw[i])))
748 VdV.w[i] = fMPY8SS(fGETBYTE(0, VuV.w[i]), fGETBYTE(0, VvV.w[i]));
749 VdV.w[i] += fMPY8SS(fGETBYTE(1, VuV.w[i]), fGETBYTE(1, VvV.w[i]));
750 VdV.w[i] += fMPY8SS(fGETBYTE(2, VuV.w[i]), fGETBYTE(2, VvV.w[i]));
751 VdV.w[i] += fMPY8SS(fGETBYTE(3, VuV.w[i]), fGETBYTE(3, VvV.w[i])))
756 VxV.w[i] += fMPY8SS(fGETBYTE(0, VuV.w[i]), fGETBYTE(0, VvV.w[i]));
757 VxV.w[i] += fMPY8SS(fGETBYTE(1, VuV.w[i]), fGETBYTE(1, VvV.w[i]));
758 VxV.w[i] += fMPY8SS(fGETBYTE(2, VuV.w[i]), fGETBYTE(2, VvV.w[i]));
759 VxV.w[i] += fMPY8SS(fGETBYTE(3, VuV.w[i]), fGETBYTE(3, VvV.w[i])))
764 VddV.v[0].uw[i] = fMPY8UU(fGETUBYTE(0, VuuV.v[uiV ? 1:0].uw[i]),fGETUBYTE((0-uiV) & 0x3,RtV));
765 VddV.v[0].uw[i] += fMPY8UU(fGETUBYTE(1, VuuV.v[0 ].uw[i]),fGETUBYTE((1-uiV) & 0x3,RtV));
766 VddV.v[0].uw[i] += fMPY8UU(fGETUBYTE(2, VuuV.v[0 ].uw[i]),fGETUBYTE((2-uiV) & 0x3,RtV));
767 VddV.v[0].uw[i] += fMPY8UU(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETUBYTE((3-uiV) & 0x3,RtV));
769 VddV.v[1].uw[i] = fMPY8UU(fGETUBYTE(0, VuuV.v[1 ].uw[i]),fGETUBYTE((2-uiV) & 0x3,RtV));
770 VddV.v[1].uw[i] += fMPY8UU(fGETUBYTE(1, VuuV.v[1 ].uw[i]),fGETUBYTE((3-uiV) & 0x3,RtV));
771 VddV.v[1].uw[i] += fMPY8UU(fGETUBYTE(2, VuuV.v[uiV ? 1:0].uw[i]),fGETUBYTE((0-uiV) & 0x3,RtV));
772 VddV.v[1].uw[i] += fMPY8UU(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETUBYTE((1-uiV) & 0x3,RtV)))
777 VxxV.v[0].uw[i] += fMPY8UU(fGETUBYTE(0, VuuV.v[uiV ? 1:0].uw[i]),fGETUBYTE((0-uiV) & 0x3,RtV));
778 VxxV.v[0].uw[i] += fMPY8UU(fGETUBYTE(1, VuuV.v[0 ].uw[i]),fGETUBYTE((1-uiV) & 0x3,RtV));
779 VxxV.v[0].uw[i] += fMPY8UU(fGETUBYTE(2, VuuV.v[0 ].uw[i]),fGETUBYTE((2-uiV) & 0x3,RtV));
780 VxxV.v[0].uw[i] += fMPY8UU(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETUBYTE((3-uiV) & 0x3,RtV));
782 VxxV.v[1].uw[i] += fMPY8UU(fGETUBYTE(0, VuuV.v[1 ].uw[i]),fGETUBYTE((2-uiV) & 0x3,RtV));
783 VxxV.v[1].uw[i] += fMPY8UU(fGETUBYTE(1, VuuV.v[1 ].uw[i]),fGETUBYTE((3-uiV) & 0x3,RtV));
784 VxxV.v[1].uw[i] += fMPY8UU(fGETUBYTE(2, VuuV.v[uiV ? 1:0].uw[i]),fGETUBYTE((0-uiV) & 0x3,RtV));
785 VxxV.v[1].uw[i] += fMPY8UU(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETUBYTE((1-uiV) & 0x3,RtV)))
796 VdV.w[i] = fMPY8US(fGETUBYTE(0,VuV.uw[i]), fGETBYTE(0,RtV));
797 VdV.w[i] += fMPY8US(fGETUBYTE(1,VuV.uw[i]), fGETBYTE(1,RtV));
798 VdV.w[i] += fMPY8US(fGETUBYTE(2,VuV.uw[i]), fGETBYTE(2,RtV));
799 VdV.w[i] += fMPY8US(fGETUBYTE(3,VuV.uw[i]), fGETBYTE(3,RtV)))
804 VxV.w[i] += fMPY8US(fGETUBYTE(0,VuV.uw[i]), fGETBYTE(0,RtV));
805 VxV.w[i] += fMPY8US(fGETUBYTE(1,VuV.uw[i]), fGETBYTE(1,RtV));
806 VxV.w[i] += fMPY8US(fGETUBYTE(2,VuV.uw[i]), fGETBYTE(2,RtV));
807 VxV.w[i] += fMPY8US(fGETUBYTE(3,VuV.uw[i]), fGETBYTE(3,RtV)))
812 VddV.v[0].w[i] = fMPY8US(fGETUBYTE(0, VuuV.v[uiV ? 1:0].uw[i]),fGETBYTE((0-uiV) & 0x3,RtV));
813 VddV.v[0].w[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0 ].uw[i]),fGETBYTE((1-uiV) & 0x3,RtV));
814 VddV.v[0].w[i] += fMPY8US(fGETUBYTE(2, VuuV.v[0 ].uw[i]),fGETBYTE((2-uiV) & 0x3,RtV));
815 VddV.v[0].w[i] += fMPY8US(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETBYTE((3-uiV) & 0x3,RtV));
817 VddV.v[1].w[i] = fMPY8US(fGETUBYTE(0, VuuV.v[1 ].uw[i]),fGETBYTE((2-uiV) & 0x3,RtV));
818 VddV.v[1].w[i] += fMPY8US(fGETUBYTE(1, VuuV.v[1 ].uw[i]),fGETBYTE((3-uiV) & 0x3,RtV));
819 VddV.v[1].w[i] += fMPY8US(fGETUBYTE(2, VuuV.v[uiV ? 1:0].uw[i]),fGETBYTE((0-uiV) & 0x3,RtV));
820 VddV.v[1].w[i] += fMPY8US(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETBYTE((1-uiV) & 0x3,RtV)))
825 VxxV.v[0].w[i] += fMPY8US(fGETUBYTE(0, VuuV.v[uiV ? 1:0].uw[i]),fGETBYTE((0-uiV) & 0x3,RtV));
826 VxxV.v[0].w[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0 ].uw[i]),fGETBYTE((1-uiV) & 0x3,RtV));
827 VxxV.v[0].w[i] += fMPY8US(fGETUBYTE(2, VuuV.v[0 ].uw[i]),fGETBYTE((2-uiV) & 0x3,RtV));
828 VxxV.v[0].w[i] += fMPY8US(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETBYTE((3-uiV) & 0x3,RtV));
830 VxxV.v[1].w[i] += fMPY8US(fGETUBYTE(0, VuuV.v[1 ].uw[i]),fGETBYTE((2-uiV) & 0x3,RtV));
831 VxxV.v[1].w[i] += fMPY8US(fGETUBYTE(1, VuuV.v[1 ].uw[i]),fGETBYTE((3-uiV) & 0x3,RtV));
832 VxxV.v[1].w[i] += fMPY8US(fGETUBYTE(2, VuuV.v[uiV ? 1:0].uw[i]),fGETBYTE((0-uiV) & 0x3,RtV));
833 VxxV.v[1].w[i] += fMPY8US(fGETUBYTE(3, VuuV.v[0 ].uw[i]),fGETBYTE((1-uiV) & 0x3,RtV)))
840 VdV.w[i] = fMPY8US(fGETUBYTE(0,VuV.uw[i]), fGETBYTE(0,VvV.w[i]));
841 VdV.w[i] += fMPY8US(fGETUBYTE(1,VuV.uw[i]), fGETBYTE(1,VvV.w[i]));
842 VdV.w[i] += fMPY8US(fGETUBYTE(2,VuV.uw[i]), fGETBYTE(2,VvV.w[i]));
843 VdV.w[i] += fMPY8US(fGETUBYTE(3,VuV.uw[i]), fGETBYTE(3,VvV.w[i])))
848 VxV.w[i] += fMPY8US(fGETUBYTE(0,VuV.uw[i]), fGETBYTE(0,VvV.w[i]));
849 VxV.w[i] += fMPY8US(fGETUBYTE(1,VuV.uw[i]), fGETBYTE(1,VvV.w[i]));
850 VxV.w[i] += fMPY8US(fGETUBYTE(2,VuV.uw[i]), fGETBYTE(2,VvV.w[i]));
851 VxV.w[i] += fMPY8US(fGETUBYTE(3,VuV.uw[i]), fGETBYTE(3,VvV.w[i])))
869 VddV.v[0].uw[i] = fABS(fGETUHALF(0, VuuV.v[0].uw[i]) - fGETUHALF(0,RtV));
870 VddV.v[0].uw[i] += fABS(fGETUHALF(1, VuuV.v[0].uw[i]) - fGETUHALF(1,RtV));
871 VddV.v[1].uw[i] = fABS(fGETUHALF(1, VuuV.v[0].uw[i]) - fGETUHALF(0,RtV));
872 VddV.v[1].uw[i] += fABS(fGETUHALF(0, VuuV.v[1].uw[i]) - fGETUHALF(1,RtV)))
876 VxxV.v[0].uw[i] += fABS(fGETUHALF(0, VuuV.v[0].uw[i]) - fGETUHALF(0,RtV));
877 VxxV.v[0].uw[i] += fABS(fGETUHALF(1, VuuV.v[0].uw[i]) - fGETUHALF(1,RtV));
878 VxxV.v[1].uw[i] += fABS(fGETUHALF(1, VuuV.v[0].uw[i]) - fGETUHALF(0,RtV));
879 VxxV.v[1].uw[i] += fABS(fGETUHALF(0, VuuV.v[1].uw[i]) - fGETUHALF(1,RtV)))
892 …VddV.v[0].uw[i] = fABS(fZE8_16(fGETUBYTE(0, VuuV.v[uiV?1:0].uw[i])) - fZE8_16(fGETUBYTE((0-uiV)&3…
893 …VddV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(1, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((1-uiV)&3…
894 …VddV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(2, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((2-uiV)&3…
895 …VddV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(3, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((3-uiV)&3…
897 …VddV.v[1].uw[i] = fABS(fZE8_16(fGETUBYTE(0, VuuV.v[1 ].uw[i])) - fZE8_16(fGETUBYTE((2-uiV)&3…
898 …VddV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(1, VuuV.v[1 ].uw[i])) - fZE8_16(fGETUBYTE((3-uiV)&3…
899 …VddV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(2, VuuV.v[uiV?1:0].uw[i])) - fZE8_16(fGETUBYTE((0-uiV)&3…
900 …VddV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(3, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((1-uiV)&3…
904 …VxxV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(0, VuuV.v[uiV?1:0].uw[i])) - fZE8_16(fGETUBYTE((0-uiV)&3…
905 …VxxV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(1, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((1-uiV)&3…
906 …VxxV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(2, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((2-uiV)&3…
907 …VxxV.v[0].uw[i] += fABS(fZE8_16(fGETUBYTE(3, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((3-uiV)&3…
909 …VxxV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(0, VuuV.v[1 ].uw[i])) - fZE8_16(fGETUBYTE((2-uiV)&3…
910 …VxxV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(1, VuuV.v[1 ].uw[i])) - fZE8_16(fGETUBYTE((3-uiV)&3…
911 …VxxV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(2, VuuV.v[uiV?1:0].uw[i])) - fZE8_16(fGETUBYTE((0-uiV)&3…
912 …VxxV.v[1].uw[i] += fABS(fZE8_16(fGETUBYTE(3, VuuV.v[0 ].uw[i])) - fZE8_16(fGETUBYTE((1-uiV)&3…
929 …t32)", "Vector arithmetic shift right " DESC, VdV.TYPE[i] = (VuV.TYPE[i] >> (RtV…
930 …t32)", "Vector arithmetic shift left " DESC, VdV.TYPE[i] = (VuV.TYPE[i] << (RtV…
931 …32)", "Vector logical shift right " DESC, VdV.u##TYPE[i] = (VuV.u##TYPE[i] >> (RtV & …
932 …ithmetic shift right " DESC, VdV.TYPE[i] = fBIDIR_ASHIFTR(VuV.TYPE[i], fSXTN((LOGSIZE+1),SI…
933 …ithmetic shift left " DESC, VdV.TYPE[i] = fBIDIR_ASHIFTL(VuV.TYPE[i], fSXTN((LOGSIZE+1),S…
934 …al shift right " DESC, VdV.u##TYPE[i] = fBIDIR_LSHIFTR(VuV.u##TYPE[i], fSXTN((LOGSIZE+1),SI…
939 …vlsrb,"Vd32.ub=vlsr(Vu32.ub,Rt32)","vec log shift right bytes", VdV.b[i] = VuV.ub[i] >> (RtV & 0x7…
941 ….uw)","Vector word rotate right", VdV.uw[i] = ((VuV.uw[i] >> (VvV.uw[i] & 0x1f)) | (VuV.uw[i] << (…
948 fHIDE(int64_t ) shift = (fSE32_64(VuV.w[i]) << 32);
949 fHIDE(int64_t ) mask = (((fSE32_64(VxxV.v[0].w[i])) << 32) | fZE32_64(VxxV.v[0].w[i]));
951 fHIDE(int ) count = -(0x40 & VvV.w[i]) + (VvV.w[i] & 0x3f);
953 VxxV.v[1].w[i] = ((result >> 32) & 0xffffffff);
954 VxxV.v[0].w[i] = (result & 0xffffffff))
964 DSTM(0,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VvV.SRCTYPE[i],shamt) >> shamt)); \
965 DSTM(1,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VuV.SRCTYPE[i],shamt) >> shamt)))
990 fHIDE(int )shamt = VvV.SRCTYPE2[2*i+0] & SHAMTMASK; \
991 DSTM(0,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VuuV.v[0].SRCTYPE[i],shamt) >> shamt)); \
992 shamt = VvV.SRCTYPE2[2*i+1] & SHAMTMASK; \
993 DSTM(1,VdV.SRCTYPE[i],SATFUNC(RNDFUNC(VuuV.v[1].SRCTYPE[i],shamt) >> shamt)))
1008 fSETHALF(0,VdV.w[i], (VvV.w[i] >> (RtV & 0xF)));
1009 fSETHALF(1,VdV.w[i], (VuV.w[i] >> (RtV & 0xF))))
1014 fSETHALF(0,VdV.w[i], fVSATH(VvV.w[i] >> (RtV & 0xF)));
1015 fSETHALF(1,VdV.w[i], fVSATH(VuV.w[i] >> (RtV & 0xF))))
1020 fSETHALF(0,VdV.w[i], fVSATH( (VvV.w[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt));
1021 fSETHALF(1,VdV.w[i], fVSATH( (VuV.w[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt)))
1026 fSETHALF(0,VdV.w[i], fVSATUH( (VvV.w[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt));
1027 fSETHALF(1,VdV.w[i], fVSATUH( (VuV.w[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt)))
1031 fSETHALF(0, VdV.uw[i], fVSATUH(VvV.w[i] >> (RtV & 0xF)));
1032 fSETHALF(1, VdV.uw[i], fVSATUH(VuV.w[i] >> (RtV & 0xF))))
1037 fSETHALF(0,VdV.w[i], fVSATUH( (VvV.uw[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt));
1038 fSETHALF(1,VdV.w[i], fVSATUH( (VuV.uw[i] + fBIDIR_ASHIFTL(1,(shamt-1),4_8) ) >> shamt)))
1045 fSETHALF(0, VdV.uw[i], fVSATH((VvV.w[i] + fCONSTLL(0x8000)) >> 16));
1046 fSETHALF(1, VdV.uw[i], fVSATH((VuV.w[i] + fCONSTLL(0x8000)) >> 16)))
1050 fSETHALF(0, VdV.uw[i], fVSATUH((VvV.w[i] + fCONSTLL(0x8000)) >> 16));
1051 fSETHALF(1, VdV.uw[i], fVSATUH((VuV.w[i] + fCONSTLL(0x8000)) >> 16)))
1055 fSETHALF(0, VdV.uw[i], fVSATUH((VvV.uw[i] + fCONSTLL(0x8000)) >> 16));
1056 fSETHALF(1, VdV.uw[i], fVSATUH((VuV.uw[i] + fCONSTLL(0x8000)) >> 16)))
1066 fSETBYTE(0, VdV.uh[i], fVSATB((VvV.h[i] + 0x80) >> 8));
1067 fSETBYTE(1, VdV.uh[i], fVSATB((VuV.h[i] + 0x80) >> 8)))
1071 fSETBYTE(0, VdV.uh[i], fVSATUB((VvV.h[i] + 0x80) >> 8));
1072 fSETBYTE(1, VdV.uh[i], fVSATUB((VuV.h[i] + 0x80) >> 8)))
1076 fSETBYTE(0, VdV.uh[i], fVSATUB((VvV.uh[i] + 0x80) >> 8));
1077 fSETBYTE(1, VdV.uh[i], fVSATUB((VuV.uh[i] + 0x80) >> 8)))
1082 VxV.w[i] += (VuV.w[i] << (RtV & (32-1))))
1086 VxV.w[i] += (VuV.w[i] >> (RtV & (32-1))))
1090 VxV.h[i] += (VuV.h[i] << (RtV & (16-1))))
1094 VxV.h[i] += (VuV.h[i] >> (RtV & (16-1))))
1108 … of Difference "DESCR, VdV.DEST[i] = (VuV.SRC[i] > VvV.SRC[i]) ? (VuV.SRC[i] - VvV.SRC[i]) : (Vv…
1111 … "Vector Add & Saturate "DESCR, VdV.DEST[i] = fVUADDSAT(WIDTH, VuV.SRC[i], VvV.SRC[i…
1112 …dV.v[0].DEST[i] = fVUADDSAT(WIDTH, VuuV.v[0].SRC[i],VvvV.v[0].SRC[i]); VddV.v[1].DEST[i] = fVUADDS…
1113 … "Vector Add & Saturate "DESCR, VdV.DEST[i] = fVUSUBSAT(WIDTH, VuV.SRC[i], VvV.SRC[i…
1114 …dV.v[0].DEST[i] = fVUSUBSAT(WIDTH, VuuV.v[0].SRC[i],VvvV.v[0].SRC[i]); VddV.v[1].DEST[i] = fVUSUBS…
1117 … "Vector Add & Saturate "DESCR, VdV.DEST[i] = fVSADDSAT(WIDTH, VuV.SRC[i], VvV.SRC[…
1118 …V.v[0].DEST[i] = fVSADDSAT(WIDTH, VuuV.v[0].SRC[i], VvvV.v[0].SRC[i]); VddV.v[1].DEST[i] = fVSADDS…
1119 … "Vector Add & Saturate "DESCR, VdV.DEST[i] = fVSSUBSAT(WIDTH, VuV.SRC[i], VvV.SRC[…
1120 …V.v[0].DEST[i] = fVSSUBSAT(WIDTH, VuuV.v[0].SRC[i], VvvV.v[0].SRC[i]); VddV.v[1].DEST[i] = fVSSUBS…
1123 …"DESCR, VdV.DEST[i] = fVAVGU( WIDTH, VuV.SRC[i], VvV.SRC[i…
1124 …% Round"DESCR, VdV.DEST[i] = fVAVGURND(WIDTH, VuV.SRC[i], VvV.SRC[i…
1129 …ESCR, VdV.DEST[i] = fVAVGS( WIDTH, VuV.SRC[i], VvV.SR…
1130 …Round"DESCR, VdV.DEST[i] = fVAVGSRND( WIDTH, VuV.SRC[i], VvV.SR…
1131 …verage "DESCR, VdV.DEST[i] = fVNAVGS( WIDTH, VuV.SRC[i], VvV.SR…
1140 …u32."#SRC",Vv32."#SRC")", "Vector Add "DESCR, VdV.DEST[i] = VuV.SRC[i] + VvV.SRC[i])\
1141 …u32."#SRC",Vv32."#SRC")", "Vector Sub "DESCR, VdV.DEST[i] = VuV.SRC[i] - VvV.SRC[i])\
1142 …dd "DESCR, VddV.v[0].DEST[i] = VuuV.v[0].SRC[i] + VvvV.v[0].SRC[i]; VddV.v[1].DEST[i] = VuuV.v[1…
1143 …ub "DESCR, VddV.v[0].DEST[i] = VuuV.v[0].SRC[i] - VvvV.v[0].SRC[i]; VddV.v[1].DEST[i] = VuuV.v[1…
1179 "Vector Negative Average Unsigned Byte", VdV.b[i] = fVNAVGU(8, VuV.ub[i], VvV.ub[i]))
1182 VdV.w[i] = fVSATW(VuV.w[i]+VvV.w[i]+fGETQBIT(QsV,i*4)))
1185 VdV.w[i] = VuV.w[i]+VvV.w[i]+fGETQBIT(QxV,i*4);
1186 fSETQBITS(QxV,4,0xF,4*i,-fCARRY_FROM_ADD32(VuV.w[i],VvV.w[i],fGETQBIT(QxV,i*4))))
1189 VdV.w[i] = VuV.w[i]+~VvV.w[i]+fGETQBIT(QxV,i*4);
1190 fSETQBITS(QxV,4,0xF,4*i,-fCARRY_FROM_ADD32(VuV.w[i],~VvV.w[i],fGETQBIT(QxV,i*4))))
1193 VdV.w[i] = VuV.w[i]+VvV.w[i];
1194 fSETQBITS(QeV,4,0xF,4*i,-fCARRY_FROM_ADD32(VuV.w[i],VvV.w[i],0)))
1197 VdV.w[i] = VuV.w[i]+~VvV.w[i]+1;
1198 fSETQBITS(QeV,4,0xF,4*i,-fCARRY_FROM_ADD32(VuV.w[i],~VvV.w[i],1)))
1201 … 64-bits (higher 32-bits come from first vector) to 32-bits",VdV.w[i] = fVSATDW(VuV.w[i],VvV.w[i]))
1205 …#SRC1",Vv32."#SRC2"):sat", "Vector Add mixed", VdV.DEST[i] = SATF(VuV.SRC1[i] + VvV.SRC2[i]))\
1206 …#SRC1",Vv32."#SRC2"):sat", "Vector Sub mixed", VdV.DEST[i] = SATF(VuV.SRC1[i] - VvV.SRC2[i]))\
1219 VddV.v[0].h[i] = fZE8_16(fGETUBYTE(0, VuV.uh[i])) + fZE8_16(fGETUBYTE(0, VvV.uh[i]));
1220 VddV.v[1].h[i] = fZE8_16(fGETUBYTE(1, VuV.uh[i])) + fZE8_16(fGETUBYTE(1, VvV.uh[i])))
1224 VddV.v[0].h[i] = fZE8_16(fGETUBYTE(0, VuV.uh[i])) - fZE8_16(fGETUBYTE(0, VvV.uh[i]));
1225 VddV.v[1].h[i] = fZE8_16(fGETUBYTE(1, VuV.uh[i])) - fZE8_16(fGETUBYTE(1, VvV.uh[i])))
1231 VddV.v[0].w[i] = fGETHALF(0, VuV.w[i]) + fGETHALF(0, VvV.w[i]);
1232 VddV.v[1].w[i] = fGETHALF(1, VuV.w[i]) + fGETHALF(1, VvV.w[i]))
1236 VddV.v[0].w[i] = fGETHALF(0, VuV.w[i]) - fGETHALF(0, VvV.w[i]);
1237 VddV.v[1].w[i] = fGETHALF(1, VuV.w[i]) - fGETHALF(1, VvV.w[i]))
1242 VddV.v[0].w[i] = fZE16_32(fGETUHALF(0, VuV.uw[i])) + fZE16_32(fGETUHALF(0, VvV.uw[i]));
1243 VddV.v[1].w[i] = fZE16_32(fGETUHALF(1, VuV.uw[i])) + fZE16_32(fGETUHALF(1, VvV.uw[i])))
1247 VddV.v[0].w[i] = fZE16_32(fGETUHALF(0, VuV.uw[i])) - fZE16_32(fGETUHALF(0, VvV.uw[i]));
1248 VddV.v[1].w[i] = fZE16_32(fGETUHALF(1, VuV.uw[i])) - fZE16_32(fGETUHALF(1, VvV.uw[i])))
1254 VxxV.v[0].w[i] += fGETHALF(0, VuV.w[i]) + fGETHALF(0, VvV.w[i]);
1255 VxxV.v[1].w[i] += fGETHALF(1, VuV.w[i]) + fGETHALF(1, VvV.w[i]))
1259 VxxV.v[0].w[i] += fGETUHALF(0, VuV.w[i]) + fGETUHALF(0, VvV.w[i]);
1260 VxxV.v[1].w[i] += fGETUHALF(1, VuV.w[i]) + fGETUHALF(1, VvV.w[i]))
1264 VxxV.v[0].h[i] += fGETUBYTE(0, VuV.h[i]) + fGETUBYTE(0, VvV.h[i]);
1265 VxxV.v[1].h[i] += fGETUBYTE(1, VuV.h[i]) + fGETUBYTE(1, VvV.h[i]))
1273 … "LHSYN"+="RHSYN,"if (Qv4) "LHSYN"+="RHSYN,DESCR,LHBEH=fCONDMASK##WIDTH(QvV,i,LHBEH+RHBEH,LHBEH)) \
1274 … "LHSYN"-="RHSYN,"if (Qv4) "LHSYN"-="RHSYN,DESCR,LHBEH=fCONDMASK##WIDTH(QvV,i,LHBEH-RHBEH,LHBEH)) \
1275 …"LHSYN"+="RHSYN,"if (!Qv4) "LHSYN"+="RHSYN,DESCR,LHBEH=fCONDMASK##WIDTH(QvV,i,LHBEH,LHBEH+RHBEH)) \
1276 …"LHSYN"-="RHSYN,"if (!Qv4) "LHSYN"-="RHSYN,DESCR,LHBEH=fCONDMASK##WIDTH(QvV,i,LHBEH,LHBEH-RHBEH)) \
1278 CONDADDSUB(8,b,"Vx32.b","Vu32.b","Conditional add/sub Byte",VxV.ub[i],VuV.ub[i])
1279 CONDADDSUB(16,h,"Vx32.h","Vu32.h","Conditional add/sub Half",VxV.h[i],VuV.h[i])
1280 CONDADDSUB(32,w,"Vx32.w","Vu32.w","Conditional add/sub Word",VxV.w[i],VuV.w[i])
1286 …", "Vd32.b=vabs(Vu32.b)", "Vector absolute value of bytes", VdV.b[i] = fABS(VuV.b[i]))
1287 …bs(Vu32.b):sat", "Vector absolute value of bytes", VdV.b[i] = fVSATB(fABS(fSE8_16(VuV.b[i]))))
1290 … "Vd32.h=vabs(Vu32.h)", "Vector absolute value of halfwords", VdV.h[i] = fABS(VuV.h[i]))
1291 …32.h):sat", "Vector absolute value of halfwords", VdV.h[i] = fVSATH(fABS(fSE16_32(VuV.h[i]))))
1292 … "Vd32.w=vabs(Vu32.w)", "Vector absolute value of words", VdV.w[i] = fABS(VuV.w[i]))
1293 …32.w):sat", "Vector absolute value of words", VdV.w[i] = fVSATW(fABS(fSE32_64(VuV.w[i]))))
1304 VddV.v[0].h[i] = fMPY8SS(fGETBYTE(0, VuV.h[i]), fGETBYTE(0, VvV.h[i]));
1305 VddV.v[1].h[i] = fMPY8SS(fGETBYTE(1, VuV.h[i]), fGETBYTE(1, VvV.h[i])))
1309 VxxV.v[0].h[i] += fMPY8SS(fGETBYTE(0, VuV.h[i]), fGETBYTE(0, VvV.h[i]));
1310 VxxV.v[1].h[i] += fMPY8SS(fGETBYTE(1, VuV.h[i]), fGETBYTE(1, VvV.h[i])))
1315 VddV.v[0].uh[i] = fMPY8UU(fGETUBYTE(0, VuV.uh[i]), fGETUBYTE(0, VvV.uh[i]) );
1316 VddV.v[1].uh[i] = fMPY8UU(fGETUBYTE(1, VuV.uh[i]), fGETUBYTE(1, VvV.uh[i]) ))
1320 VxxV.v[0].uh[i] += fMPY8UU(fGETUBYTE(0, VuV.uh[i]), fGETUBYTE(0, VvV.uh[i]) );
1321 VxxV.v[1].uh[i] += fMPY8UU(fGETUBYTE(1, VuV.uh[i]), fGETUBYTE(1, VvV.uh[i]) ))
1330 VddV.v[0].h[i] = fMPY8US(fGETUBYTE(0, VuV.uh[i]), fGETBYTE(0, VvV.h[i]));
1331 VddV.v[1].h[i] = fMPY8US(fGETUBYTE(1, VuV.uh[i]), fGETBYTE(1, VvV.h[i])))
1335 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(0, VuV.uh[i]), fGETBYTE(0, VvV.h[i]));
1336 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(1, VuV.uh[i]), fGETBYTE(1, VvV.h[i])))
1343 …[0].h[i] = fMPY8US(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETBYTE(0, VvvV.v[0].uh[i])) + fMPY8US(fGETUBYT…
1344 …[1].h[i] = fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETBYTE(1, VvvV.v[0].uh[i])) + fMPY8US(fGETUBYT…
1348 …0].h[i] = fMPY8UU(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETUBYTE(0, VvvV.v[0].uh[i])) + fMPY8UU(fGETUBYT…
1349 …1].h[i] = fMPY8UU(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETUBYTE(1, VvvV.v[0].uh[i])) + fMPY8UU(fGETUBYT…
1359 VddV.v[0].w[i] = fMPY16SS(fGETHALF(0, VuV.w[i]), fGETHALF(0, VvV.w[i]));
1360 VddV.v[1].w[i] = fMPY16SS(fGETHALF(1, VuV.w[i]), fGETHALF(1, VvV.w[i])))
1364 VxxV.v[0].w[i] += fMPY16SS(fGETHALF(0, VuV.w[i]), fGETHALF(0, VvV.w[i]));
1365 VxxV.v[1].w[i] += fMPY16SS(fGETHALF(1, VuV.w[i]), fGETHALF(1, VvV.w[i])))
1369 VddV.v[0].uw[i] = fMPY16UU(fGETUHALF(0, VuV.uw[i]), fGETUHALF(0, VvV.uw[i]));
1370 VddV.v[1].uw[i] = fMPY16UU(fGETUHALF(1, VuV.uw[i]), fGETUHALF(1, VvV.uw[i])))
1374 VxxV.v[0].uw[i] += fMPY16UU(fGETUHALF(0, VuV.uw[i]), fGETUHALF(0, VvV.uw[i]));
1375 VxxV.v[1].uw[i] += fMPY16UU(fGETUHALF(1, VuV.uw[i]), fGETUHALF(1, VvV.uw[i])))
1382 VdV.h[i] = fVSATH(fGETHALF(1,fVSAT(fROUND((fMPY16SS(VuV.h[i],VvV.h[i] )<<1))))))
1388 VdV.uh[i] = fGETUHALF(1,fMPY16UU(VuV.uh[i],VvV.uh[i])))
1393 VddV.v[0].w[i] = fMPY16SU(fGETHALF(0, VuV.w[i]), fGETUHALF(0, VvV.uw[i]));
1394 VddV.v[1].w[i] = fMPY16SU(fGETHALF(1, VuV.w[i]), fGETUHALF(1, VvV.uw[i])))
1399 VxxV.v[0].w[i] += fMPY16SU(fGETHALF(0, VuV.w[i]), fGETUHALF(0, VvV.uw[i]));
1400 VxxV.v[1].w[i] += fMPY16SU(fGETHALF(1, VuV.w[i]), fGETUHALF(1, VvV.uw[i])))
1407 VdV.h[i] = fMPY16SS(VuV.h[i], VvV.h[i]))
1411 VxV.h[i] += fMPY16SS(VuV.h[i], VvV.h[i]))
1420 VdV.w[i] = fMPY3216SU(VuV.w[i], fGETUHALF(0, VvV.w[i])) >> 16)
1424 VdV.w[i] = fVSATW((((fMPY3216SS(VuV.w[i], fGETHALF(1, VvV.w[i])) >> 14) + 0) >> 1)))
1428 VdV.w[i] = fVSATW((((fMPY3216SS(VuV.w[i], fGETHALF(1, VvV.w[i])) >> 14) + 1) >> 1)))
1433 prod = fMPY32SU(VuV.w[i],fGETUHALF(0,VvV.w[i]));
1434 VddV.v[1].w[i] = prod >> 16;
1435 VddV.v[0].w[i] = prod << 16)
1440 prod = fMPY32SS(VuV.w[i],fGETHALF(1,VvV.w[i])) + fSE32_64(VxxV.v[1].w[i]);
1441 VxxV.v[1].w[i] = prod >> 16;
1442 fSETHALF(0, VxxV.v[0].w[i], VxxV.v[0].w[i] >> 16);
1443 fSETHALF(1, VxxV.v[0].w[i], prod & 0x0000ffff))
1447 IV1DEAD() VxV.w[i] = fVSATW(((((VxV.w[i] + fMPY3216SS(VuV.w[i], fGETHALF(1, VvV.w[i]))) >> 14) + 0)…
1451 IV1DEAD() VxV.w[i] = fVSATW(((((VxV.w[i] + fMPY3216SS(VuV.w[i], fGETHALF(1, VvV.w[i]))) >> 14) + 1)…
1456 VdV.w[i] = (fGETHALF(0,VuV.w[i])*fGETHALF(1,VvV.w[i])) << 16)
1460 IV1DEAD() VdV.w[i] = fMPY3216SU(VuV.w[i], fGETUHALF(0, VvV.w[i])) )
1464 IV1DEAD() VdV.w[i] = fMPY3216SS(VuV.w[i], fGETHALF(1, VvV.w[i])) )
1470 VxV.w[i] = VxV.w[i] + fMPY3216SS(VuV.w[i], fGETHALF(0, VvV.w[i])) )
1474 VxV.w[i] = VxV.w[i] + fMPY3216SU(VuV.w[i], fGETUHALF(0, VvV.w[i])) )
1485 VddV.v[0].uh[i] = fMPY8UU(fGETUBYTE(0, VuV.uh[i]), fGETUBYTE((2*i+0)%4, RtV));
1486 VddV.v[1].uh[i] = fMPY8UU(fGETUBYTE(1, VuV.uh[i]), fGETUBYTE((2*i+1)%4, RtV)))
1490 VxxV.v[0].uh[i] += fMPY8UU(fGETUBYTE(0, VuV.uh[i]), fGETUBYTE((2*i+0)%4, RtV));
1491 VxxV.v[1].uh[i] += fMPY8UU(fGETUBYTE(1, VuV.uh[i]), fGETUBYTE((2*i+1)%4, RtV)))
1496 VddV.v[0].h[i] = fMPY8US(fGETUBYTE(0, VuV.uh[i]), fGETBYTE((2*i+0)%4, RtV));
1497 VddV.v[1].h[i] = fMPY8US(fGETUBYTE(1, VuV.uh[i]), fGETBYTE((2*i+1)%4, RtV)))
1501 VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(0, VuV.uh[i]), fGETBYTE((2*i+0)%4, RtV));
1502 VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(1, VuV.uh[i]), fGETBYTE((2*i+1)%4, RtV)))
1507 …VddV.v[0].h[i] = fMPY8US(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETBYTE(0, RtV)) + fMPY16SS(fGETUBYTE(0, …
1508 …VddV.v[1].h[i] = fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETBYTE(2, RtV)) + fMPY16SS(fGETUBYTE(1, …
1512 …VxxV.v[0].h[i] += fMPY8US(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETBYTE(0, RtV)) + fMPY16SS(fGETUBYTE(0,…
1513 …VxxV.v[1].h[i] += fMPY8US(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETBYTE(2, RtV)) + fMPY16SS(fGETUBYTE(1,…
1519 …VddV.v[0].uh[i] = fMPY8UU(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETUBYTE(0, RtV)) + fMPY8UU(fGETUBYTE(0,…
1520 …VddV.v[1].uh[i] = fMPY8UU(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETUBYTE(2, RtV)) + fMPY8UU(fGETUBYTE(1,…
1524 …VxxV.v[0].uh[i] += fMPY8UU(fGETUBYTE(0, VuuV.v[0].uh[i]), fGETUBYTE(0, RtV)) + fMPY8UU(fGETUBYTE(0…
1525 …VxxV.v[1].uh[i] += fMPY8UU(fGETUBYTE(1, VuuV.v[0].uh[i]), fGETUBYTE(2, RtV)) + fMPY8UU(fGETUBYTE(1…
1533 …VddV.v[0].w[i] = fMPY16SS(fGETHALF(0, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(0, RtV))) + fMPY16SS(fGETH…
1534 …VddV.v[1].w[i] = fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(2, RtV))) + fMPY16SS(fGETH…
1538 …VxxV.v[0].w[i] += fMPY16SS(fGETHALF(0, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(0, RtV))) + fMPY16SS(fGET…
1539 …VxxV.v[1].w[i] += fMPY16SS(fGETHALF(1, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(2, RtV))) + fMPY16SS(fGET…
1544 …VddV.v[0].w[i] = fMPY16US(fGETUHALF(0, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(0, RtV))) + fMPY16US(fGET…
1545 …VddV.v[1].w[i] = fMPY16US(fGETUHALF(1, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(2, RtV))) + fMPY16US(fGET…
1549 …VxxV.v[0].w[i] += fMPY16US(fGETUHALF(0, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(0, RtV))) + fMPY16US(fGE…
1550 …VxxV.v[1].w[i] += fMPY16US(fGETUHALF(1, VuuV.v[0].w[i]), fSE8_16(fGETBYTE(2, RtV))) + fMPY16US(fGE…
1561 VddV.v[0].w[i] = fMPY16SS(fGETHALF(0, VuV.w[i]), fGETHALF(0, RtV));
1562 VddV.v[1].w[i] = fMPY16SS(fGETHALF(1, VuV.w[i]), fGETHALF(1, RtV)))
1566 VxxV.v[0].w[i] = fCAST8s(VxxV.v[0].w[i]) + fMPY16SS(fGETHALF(0, VuV.w[i]), fGETHALF(0, RtV));
1567 VxxV.v[1].w[i] = fCAST8s(VxxV.v[1].w[i]) + fMPY16SS(fGETHALF(1, VuV.w[i]), fGETHALF(1, RtV)))
1572 …VxxV.v[0].w[i] = fVSATW(fCAST8s(VxxV.v[0].w[i]) + fMPY16SS(fGETHALF(0, VuV.w[i]), fGETHALF(0, RtV…
1573 …VxxV.v[1].w[i] = fVSATW(fCAST8s(VxxV.v[1].w[i]) + fMPY16SS(fGETHALF(1, VuV.w[i]), fGETHALF(1, RtV…
1579 …fSETHALF(0,VdV.w[i],fVSATH(fGETHALF(1,fVSAT((fMPY16SS(fGETHALF(0,VuV.w[i]),fGETHALF(0,RtV))<<1))))…
1580 …fSETHALF(1,VdV.w[i],fVSATH(fGETHALF(1,fVSAT((fMPY16SS(fGETHALF(1,VuV.w[i]),fGETHALF(1,RtV))<<1))))…
1585 …fSETHALF(0,VdV.w[i],fVSATH(fGETHALF(1,fVSAT(fROUND((fMPY16SS(fGETHALF(0,VuV.w[i]),fGETHALF(0,RtV))…
1586 …fSETHALF(1,VdV.w[i],fVSATH(fGETHALF(1,fVSAT(fROUND((fMPY16SS(fGETHALF(1,VuV.w[i]),fGETHALF(1,RtV))…
1592 VddV.v[0].uw[i] = fMPY16UU(fGETUHALF(0, VuV.uw[i]),fGETUHALF(0,RtV));
1593 VddV.v[1].uw[i] = fMPY16UU(fGETUHALF(1, VuV.uw[i]),fGETUHALF(1,RtV)))
1598 VxxV.v[0].uw[i] += fMPY16UU(fGETUHALF(0, VuV.uw[i]),fGETUHALF(0,RtV));
1599 VxxV.v[1].uw[i] += fMPY16UU(fGETUHALF(1, VuV.uw[i]),fGETUHALF(1,RtV)))
1609 VdV.h[i] = fMPY16SS(VuV.h[i], fGETBYTE(i % 4, RtV) ))
1613 VxV.h[i] += fMPY16SS(VuV.h[i], fGETBYTE(i % 4, RtV) ))
1621 VdV.w[i] = fMPY32SS(VuV.w[i], fGETBYTE(i % 4, RtV) ))
1625 VxV.w[i] += fMPY32SS(VuV.w[i], fGETBYTE(i % 4, RtV) ))
1629 VdV.w[i] = fMPY32SS(VuV.w[i], fGETUBYTE(i % 4, RtV) ))
1633 VxV.w[i] += fMPY32SS(VuV.w[i], fGETUBYTE(i % 4, RtV) ))
1641 VdV.w[i] = fMPY32SS(VuV.w[i], fGETHALF(i % 2, RtV)))
1645 VxV.w[i] += fMPY32SS(VuV.w[i], fGETHALF(i % 2, RtV)))
1668 …SN_ANY_SLOT(16,vand,"Vd32=vand(Vu32,Vv32)", "Vector Logical And", VdV.uh[i] = VuV.uh[i] & VvV.h[i])
1669 …NSN_ANY_SLOT(16,vor, "Vd32=vor(Vu32,Vv32)", "Vector Logical Or", VdV.uh[i] = VuV.uh[i] | VvV.h[i])
1670 …ANY_SLOT(16,vxor,"Vd32=vxor(Vu32,Vv32)", "Vector Logical XOR", VdV.uh[i] = VuV.uh[i] ^ VvV.h[i])
1671 ITERATOR_INSN_ANY_SLOT(16,vnot,"Vd32=vnot(Vu32)", "Vector Logical NOT", VdV.uh[i] = ~VuV.uh[i])
1679 VdV.ub[i] = fGETQBIT(QuV,i) ? fGETUBYTE(i % 4, RtV) : 0)
1683 VxV.ub[i] |= (fGETQBIT(QuV,i)) ? fGETUBYTE(i % 4, RtV) : 0)
1687 VdV.ub[i] = !fGETQBIT(QuV,i) ? fGETUBYTE(i % 4, RtV) : 0)
1691 VxV.ub[i] |= !(fGETQBIT(QuV,i)) ? fGETUBYTE(i % 4, RtV) : 0)
1696 fSETQBIT(QdV,i,((VuV.ub[i] & fGETUBYTE(i % 4, RtV)) != 0) ? 1 : 0))
1700 fSETQBIT(QxV,i,fGETQBIT(QxV,i)|(((VuV.ub[i] & fGETUBYTE(i % 4, RtV)) != 0) ? 1 : 0)))
1703 VdV.b[i] = fGETQBIT(QvV,i) ? VuV.b[i] : 0)
1705 VdV.b[i] = !fGETQBIT(QvV,i) ? VuV.b[i] : 0)
1713 for(fHIDE(int) i = 0; i < fVBYTES(); i += WIDTH) { \
1714 fSETQBITS(DEST,WIDTH,MASK,i,ASRC ASRCOP ((VuV.SRC[i/WIDTH] CMP VvV.SRC[i/WIDTH]) ? MASK : 0)); \
1723 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), &, >, N, SRC, MASK, WIDTH)) \
1725 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), |, >, N, SRC, MASK, WIDTH)) \
1727 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), ^, >, N, SRC, MASK, WIDTH))
1734 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), &, ==, N, SRC, MASK, WIDTH)) \
1736 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), |, ==, N, SRC, MASK, WIDTH)) \
1738 VCMP(QxV, fGETQBITS(QxV,WIDTH,MASK,i), ^, ==, N, SRC, MASK, WIDTH))
1754 fHIDE(int i;)
1755 for(i = 0; i < fVBYTES(); i++) fSETQBIT(QdV,i,(i < (RtV & (fVBYTES()-1))) ? 1 : 0);
1760 fHIDE(int i;)
1761 for(i = 0; i < fVBYTES(); i++) fSETQBIT(QdV,i,(i <= ((RtV-1) & (fVBYTES()-1))) ? 1 : 0);
1765 …huffe(Qs4.w,Qt4.w)","Shrink Predicate", fSETQBIT(QdV,i, (i & 2) ? fGETQBIT(QsV,i-2) : fGETQBIT(QtV…
1766 …huffe(Qs4.h,Qt4.h)","Shrink Predicate", fSETQBIT(QdV,i, (i & 1) ? fGETQBIT(QsV,i-1) : fGETQBIT(QtV…
1767 …_or, "Qd4=or(Qs4,Qt4)","Vector Predicate Or", fSETQBIT(QdV,i,fGETQBIT(QsV,i) || fGETQBIT(QtV,i) ) )
1768 …d, "Qd4=and(Qs4,Qt4)","Vector Predicate And", fSETQBIT(QdV,i,fGETQBIT(QsV,i) && fGETQBIT(QtV,i) ) )
1769 …or, "Qd4=xor(Qs4,Qt4)","Vector Predicate Xor", fSETQBIT(QdV,i,fGETQBIT(QsV,i) ^ fGETQBIT(QtV,i) ) )
1770 …r(Qs4,!Qt4)","Vector Predicate Or with not", fSETQBIT(QdV,i,fGETQBIT(QsV,i) || !fGETQBIT(QtV,i) ) )
1771 …Qs4,!Qt4)","Vector Predicate And with not", fSETQBIT(QdV,i,fGETQBIT(QsV,i) && !fGETQBIT(QtV,i) ) )
1772 …SN_ANY_SLOT(8, pred_not, "Qd4=not(Qs4)","Vector Predicate Not", fSETQBIT(QdV,i,!fGETQBIT(QsV,i) ) )
1779 fHIDE(int i;)
1780 fVFOREACH(8, i) {
1781 VdV.ub[i] = VuV.ub[i];
1789 fHIDE(int i;)
1790 fVFOREACH(8, i) {
1791 VdV.ub[i] = VuV.ub[i];
1799 fHIDE(int i;)
1800 fVFOREACH(8, i) {
1801 VddV.v[0].ub[i] = VvV.ub[i];
1802 VddV.v[1].ub[i] = VuV.ub[i];
1810 fHIDE(int i;)
1811 fVFOREACH(8, i) {
1812 VddV.v[0].ub[i] = VvV.ub[i];
1813 VddV.v[1].ub[i] = VuV.ub[i];
1822 VdV.ub[i] = fGETQBIT(QtV,i) ? VuV.ub[i] : VvV.ub[i])
1826 VddV.v[0].ub[i] = fGETQBIT(QtV,i) ? VuV.ub[i] : VvV.ub[i];
1827 VddV.v[1].ub[i] = !fGETQBIT(QtV,i) ? VuV.ub[i] : VvV.ub[i])
1837 …,Vv32."#SRC")", "Vector " DESCR " max", VdV.SRC[i] = (VuV.SRC[i] > VvV.SRC[i]) ? VuV.SRC[i] : VvV…
1838 …,Vv32."#SRC")", "Vector " DESCR " min", VdV.SRC[i] = (VuV.SRC[i] < VvV.SRC[i]) ? VuV.SRC[i] : VvV…
1860 fSETBYTE(0, VdV.uh[i], fVSATUB(VvV.h[i]));
1861 fSETBYTE(1, VdV.uh[i], fVSATUB(VuV.h[i])))
1865 fSETHALF(0, VdV.w[i], fVSATH(VvV.w[i]));
1866 fSETHALF(1, VdV.w[i], fVSATH(VuV.w[i])))
1870 fSETHALF(0, VdV.w[i], fVSATUH(VvV.uw[i]));
1871 fSETHALF(1, VdV.w[i], fVSATUH(VuV.uw[i])))
1875 fSETBYTE(0, VdV.uh[i], fGETUBYTE(0, VvV.uh[i]));
1876 fSETBYTE(1, VdV.uh[i], fGETUBYTE(0, VuV.uh[i])))
1880 fSETBYTE(0, VdV.uh[i], fGETUBYTE(1, VvV.uh[i]));
1881 fSETBYTE(1, VdV.uh[i], fGETUBYTE(1, VuV.uh[i])))
1885 fSETHALF(0, VdV.uw[i], fGETUHALF(0, VvV.uw[i]));
1886 fSETHALF(1, VdV.uw[i], fGETUHALF(0, VuV.uw[i])))
1890 fSETHALF(0, VdV.uw[i], fGETUHALF(1, VvV.uw[i]));
1891 fSETHALF(1, VdV.uw[i], fGETUHALF(1, VuV.uw[i])))
1978 fSETHALF(0, VddV.v[0].uw[i], fGETUHALF(0, VvV.uw[i]));
1979 fSETHALF(1, VddV.v[0].uw[i], fGETUHALF(0, VuV.uw[i]));
1980 fSETHALF(0, VddV.v[1].uw[i], fGETUHALF(1, VvV.uw[i]));
1981 fSETHALF(1, VddV.v[1].uw[i], fGETUHALF(1, VuV.uw[i])))
1985 fSETBYTE(0, VddV.v[0].uh[i], fGETUBYTE(0, VvV.uh[i]));
1986 fSETBYTE(1, VddV.v[0].uh[i], fGETUBYTE(0, VuV.uh[i]));
1987 fSETBYTE(0, VddV.v[1].uh[i], fGETUBYTE(1, VvV.uh[i]));
1988 fSETBYTE(1, VddV.v[1].uh[i], fGETUBYTE(1, VuV.uh[i])))
1997 VdV.uh[i ] = fGETUHALF(0, VuV.uw[i]);
1998 VdV.uh[i+fVELEM(32)] = fGETUHALF(1, VuV.uw[i]))
2002 VdV.ub[i ] = fGETUBYTE(0, VuV.uh[i]);
2003 VdV.ub[i+fVELEM(16)] = fGETUBYTE(1, VuV.uh[i]))
2007 VdV.ub[0+i ] = fGETUBYTE(0, VvV.uw[i]);
2008 VdV.ub[fVELEM(32)+i ] = fGETUBYTE(2, VvV.uw[i]);
2009 VdV.ub[2*fVELEM(32)+i] = fGETUBYTE(0, VuV.uw[i]);
2010 VdV.ub[3*fVELEM(32)+i] = fGETUBYTE(2, VuV.uw[i]))
2018 fSETHALF(0, VdV.uw[i], VuV.uh[i]);
2019 fSETHALF(1, VdV.uw[i], VuV.uh[i+fVELEM(32)]))
2023 fSETBYTE(0, VdV.uh[i], VuV.ub[i]);
2024 fSETBYTE(1, VdV.uh[i], VuV.ub[i+fVELEM(16)]))
2048 …LATE(32,lvsplatw, "Vd32=vsplat(Rt32)", "Replicates scalar across words in vector", VdV.uw[i] = RtV)
2050 …E(16,lvsplath, "Vd32.h=vsplat(Rt32)", "Replicates scalar across halves in vector", VdV.uh[i] = RtV)
2052 …ATE(8,lvsplatb, "Vd32.b=vsplat(Rt32)", "Replicates scalar across bytes in vector", VdV.ub[i] = RtV)
2055 ITERATOR_INSN_ANY_SLOT(32,vassign,"Vd32=Vu32","Copy a vector",VdV.w[i]=VuV.w[i])
2060 VddV.v[0].ub[i] = VvV.ub[i];
2061 VddV.v[1].ub[i] = VuV.ub[i])
2070 fHIDE(int i;)
2071 fVFOREACH(8, i) {
2072 VddV.v[0].ub[i] = VvV.ub[i];
2073 VddV.v[1].ub[i] = VuV.ub[i];
2080 fHIDE(int i;)
2081 fVFOREACH(32, i) {
2082 VdV.w[i]=VuV.w[i];
2130 …","Vd32.uw=vcl0(Vu32.uw)", "Count Leading Zeros in Word", VdV.uw[i]=fCL1_4(~VuV.uw[i]))
2131 …)","Vd32.uh=vcl0(Vu32.uh)", "Count Leading Zeros in Word", VdV.uh[i]=fCL1_2(~VuV.uh[i]))
2134 VdV.w[i]=fMAX(fCL1_4(~VuV.w[i]),fCL1_4(VuV.w[i]))-1; fHIDE(IV1DEAD();))
2136 VdV.h[i]=fMAX(fCL1_2(~VuV.h[i]),fCL1_2(VuV.h[i]))-1; fHIDE(IV1DEAD();))
2140 VdV.w[i] = fMAX(fCL1_4(~VuV.w[i]),fCL1_4(VuV.w[i])) + VvV.w[i])
2144 VdV.h[i] = fMAX(fCL1_2(~VuV.h[i]),fCL1_2(VuV.h[i])) + VvV.h[i])
2147 …)","Vd32.h=vpopcount(Vu32.h)", "Count Leading Zeros in Word", VdV.uh[i]=fCOUNTONES_2(VuV.uh[i]))
2155 for (fHIDE(int )i=0; i<128/8; ++i) { \
2156 unsigned char value = INPUTVEC.ub[(128/8)*lane+i]; \
2170 for (fHIDE(int )i=0; i<128/8; ++i) { \
2171 unsigned char value = INPUTVEC.ub[(128/8)*lane+i]; \
2175 if (fGETQBIT(QVAL,128/8*lane+i)) tmp.uh[(128/16)*lane+(element)]++; \
2194 fHIDE(unsigned int) bucket = fGETUBYTE(0,input.h[i]); \
2195 fHIDE(unsigned int) weight = fGETUBYTE(1,input.h[i]); \
2197 fHIDE(unsigned int) elindex = ((i>>BSHIFT) & (~MASK)) | ((bucket>>BSHIFT) & MASK); \
2205 …hist256(Qv4)","vector weighted histogram halfword counters", WHIST(uh,7,0,if (fGETQBIT(QvV,2*i)),))
2207 …):sat","vector weighted histogram halfword counters", WHIST(uh,7,0,if (fGETQBIT(QvV,2*i)),fVSATUH))
2209 …,"vwhist128(Qv4)","vector weighted histogram word counters", WHIST(uw,3,1,if (fGETQBIT(QvV,2*i)),))
2211 … weighted histogram word counters", WHIST(uw,3,1,if (((bucket & 1) == uiV) && fGETQBIT(QvV,2*i)),))
2226 idx = VuV.ub[i];
2227 VdV.b[i] = ((idx & 0xE0) == (matchval << 5)) ? fGETBYTE(oddhalf,VvV.h[idx % fVELEM(16)]) : 0)
2234 idx = VuV.ub[i];
2235 VxV.b[i] |= ((idx & 0xE0) == (matchval << 5)) ? fGETBYTE(oddhalf,VvV.h[idx % fVELEM(16)]) : 0)
2241 idx = fGETUBYTE(0,VuV.uh[i]);
2242 VddV.v[0].h[i] = ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0;
2243 idx = fGETUBYTE(1,VuV.uh[i]);
2244 VddV.v[1].h[i] = ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0)
2250 idx = fGETUBYTE(0,VuV.uh[i]);
2251 VxxV.v[0].h[i] |= ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0;
2252 idx = fGETUBYTE(1,VuV.uh[i]);
2253 VxxV.v[1].h[i] |= ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0)
2259 idx = VuV.ub[i];
2260 VdV.b[i] = ((idx & 0xE0) == (matchval << 5)) ? fGETBYTE(oddhalf,VvV.h[idx % fVELEM(16)]) : 0)
2267 idx = VuV.ub[i];
2268 VxV.b[i] |= ((idx & 0xE0) == (matchval << 5)) ? fGETBYTE(oddhalf,VvV.h[idx % fVELEM(16)]) : 0)
2274 idx = fGETUBYTE(0,VuV.uh[i]);
2275 VddV.v[0].h[i] = ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0;
2276 idx = fGETUBYTE(1,VuV.uh[i]);
2277 VddV.v[1].h[i] = ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0)
2283 idx = fGETUBYTE(0,VuV.uh[i]);
2284 VxxV.v[0].h[i] |= ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0;
2285 idx = fGETUBYTE(1,VuV.uh[i]);
2286 VxxV.v[1].h[i] |= ((idx & 0xF0) == (matchval << 4)) ? fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]) : 0)
2292 idx = VuV.ub[i];
2294 VdV.b[i] = fGETBYTE(oddhalf,VvV.h[idx % fVELEM(16)]))
2300 idx = fGETUBYTE(0,VuV.uh[i]);
2302 VddV.v[0].h[i] = fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]);
2303 idx = fGETUBYTE(1,VuV.uh[i]);
2305 VddV.v[1].h[i] = fGETHALF(oddhalf,VvV.w[idx % fVELEM(32)]))
2315 …VxV.h[i]= fVSATH( ( ( fMPY16SS(VxV.h[i],VuV.h[i])<<1) + (fGETHALF(( (VuV.h[i]>>14)&0x3), RttV )<<1…
2319 …VxV.h[i]= fVSATH( ( fMPY16SU(VxV.h[i],VuV.uh[i]) + (fGETUHALF(((VuV.uh[i]>>14)&0x3), RttV )<<15))…
2322 …VxV.h[i]= fVSATH( ( fMPY16SU(VxV.h[i],VuV.uh[i]) - (fGETUHALF(((VuV.uh[i]>>14)&0x3), RttV )<<15))…
2326 VdV.h[i]= fGETHALF( ((VuV.h[i]>>14)&0x3), RttV ))
2336 VdV.uw[i] = fMPY16UU(fGETUHALF(0, VuV.uw[i]),fGETUHALF(0,RtV)))
2341 VxV.uw[i] += fMPY16UU(fGETUHALF(0, VuV.uw[i]),fGETUHALF(0,RtV)))
2348 fHIDE(int i;)
2353 fVFOREACH(32, i) {
2354 EA = RtV+VvV.uw[i];
2355 fVLOG_VTCM_GATHER_WORD(EA, VvV.uw[i], i,MuV);
2361 fHIDE(int i;)
2366 fVFOREACH(16, i) {
2367 EA = RtV+VvV.uh[i];
2368 fVLOG_VTCM_GATHER_HALFWORD(EA, VvV.uh[i], i,MuV);
2377 fHIDE(int i;)
2383 fVFOREACH(32, i) {
2385 EA = RtV+VvvV.v[j].uw[i];
2386 fVLOG_VTCM_GATHER_HALFWORD_DV(EA, VvvV.v[j].uw[i], (2*i+j),i,j,MuV);
2395 fHIDE(int i;)
2400 fVFOREACH(32, i) {
2401 EA = RtV+VvV.uw[i];
2402 fVLOG_VTCM_GATHER_WORDQ(EA, VvV.uw[i], i,QsV,MuV);
2408 fHIDE(int i;)
2413 fVFOREACH(16, i) {
2414 EA = RtV+VvV.uh[i];
2415 fVLOG_VTCM_GATHER_HALFWORDQ(EA, VvV.uh[i], i,QsV,MuV);
2424 fHIDE(int i;)
2430 fVFOREACH(32, i) {
2432 EA = RtV+VvvV.v[j].uw[i];
2433 fVLOG_VTCM_GATHER_HALFWORDQ_DV(EA, VvvV.v[j].uw[i], (2*i+j),i,j,QsV,MuV);
2443 fHIDE(int i;)
2448 fVFOREACH(32, i) {
2449 EA = RtV+VvV.uw[i];
2450 fVLOG_VTCM_WORD(EA, VvV.uw[i], VwV,i,MuV);
2459 fHIDE(int i;)
2464 fVFOREACH(16, i) {
2465 EA = RtV+VvV.uh[i];
2466 fVLOG_VTCM_HALFWORD(EA,VvV.uh[i],VwV,i,MuV);
2474 fHIDE(int i;)
2480 fVFOREACH(32, i) {
2481 EA = (RtV+fVALIGN(VvV.uw[i],ALIGNMENT));
2482 fVLOG_VTCM_WORD_INCREMENT(EA,VvV.uw[i],VwV,i,ALIGNMENT,MuV);
2490 fHIDE(int i;)
2496 fVFOREACH(16, i) {
2497 EA = (RtV+fVALIGN(VvV.uh[i],ALIGNMENT));
2498 fVLOG_VTCM_HALFWORD_INCREMENT(EA,VvV.uh[i],VwV,i,ALIGNMENT,MuV);
2507 fHIDE(int i;)
2512 fVFOREACH(32, i) {
2513 EA = RtV+VvV.uw[i];
2514 fVLOG_VTCM_WORDQ(EA,VvV.uw[i], VwV,i,QsV,MuV);
2521 fHIDE(int i;)
2526 fVFOREACH(16, i) {
2527 EA = RtV+VvV.uh[i];
2528 fVLOG_VTCM_HALFWORDQ(EA,VvV.uh[i],VwV,i,QsV,MuV);
2538 fHIDE(int i;)
2544 fVFOREACH(32, i) {
2546 EA = RtV+VvvV.v[j].uw[i];
2547 fVLOG_VTCM_HALFWORD_DV(EA,VvvV.v[j].uw[i],VwV,(2*i+j),i,j,MuV);
2556 fGET10BIT(c00, VvvV.v[0].uw[i], 0)
2558 fGET10BIT(c01, VvvV.v[0].uw[i], 1)
2560 fGET10BIT(c02, VvvV.v[0].uw[i], 2)
2563 fGET10BIT(c10, VvvV.v[1].uw[i], 0)
2565 fGET10BIT(c11, VvvV.v[1].uw[i], 1)
2567 fGET10BIT(c12, VvvV.v[1].uw[i], 2)
2570 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c10);
2571 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c11);
2572 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c12);
2574 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c00);
2575 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2576 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c02);
2578 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c10);
2579 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c11);
2580 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c12);
2583 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c00);
2584 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c01);
2585 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c02);
2587 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c10);
2588 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c11);
2589 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c12);
2591 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c00);
2592 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2593 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c02);
2596 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c10);
2597 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2598 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c12);
2600 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c00);
2601 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c01);
2602 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c02);
2604 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c10);
2605 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c11);
2606 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c12);
2609 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c00);
2610 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c01);
2611 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c02);
2613 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c10);
2614 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2615 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c12);
2617 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c00);
2618 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c01);
2619 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c02);
2624 fGET10BIT(c00, VvvV.v[0].uw[i], 0)
2626 fGET10BIT(c01, VvvV.v[0].uw[i], 1)
2628 fGET10BIT(c02, VvvV.v[0].uw[i], 2)
2630 fGET10BIT(c10, VvvV.v[1].uw[i], 0)
2632 fGET10BIT(c11, VvvV.v[1].uw[i], 1)
2634 fGET10BIT(c12, VvvV.v[1].uw[i], 2)
2637 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c10);
2638 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c11);
2639 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c12);
2641 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c00);
2642 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2643 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c02);
2645 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c10);
2646 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c11);
2647 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c12);
2650 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c00);
2651 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c01);
2652 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c02);
2654 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c10);
2655 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c11);
2656 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c12);
2658 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c00);
2659 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2660 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c02);
2663 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c10);
2664 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2665 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c12);
2667 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c00);
2668 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c01);
2669 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c02);
2671 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c10);
2672 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c11);
2673 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c12);
2676 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c00);
2677 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c01);
2678 VxxV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c02);
2680 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c10);
2681 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2682 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c12);
2684 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c00);
2685 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c01);
2686 VxxV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c02);
2693 fGET10BIT(c00, VvvV.v[0].uw[i], 0)
2695 fGET10BIT(c01, VvvV.v[0].uw[i], 1)
2697 fGET10BIT(c02, VvvV.v[0].uw[i], 2)
2699 fGET10BIT(c10, VvvV.v[1].uw[i], 0)
2701 fGET10BIT(c11, VvvV.v[1].uw[i], 1)
2703 fGET10BIT(c12, VvvV.v[1].uw[i], 2)
2708 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c10);
2709 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c11);
2710 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c12);
2712 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c00);
2713 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2714 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c02);
2716 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c10);
2717 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c11);
2718 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c12);
2721 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c00);
2722 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c01);
2723 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c02);
2725 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c10);
2726 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c11);
2727 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c12);
2729 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c00);
2730 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2731 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c02);
2734 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c10);
2735 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2736 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c12);
2738 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c00);
2739 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c01);
2740 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c02);
2742 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c10);
2743 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c11);
2744 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c12);
2747 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c00);
2748 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c01);
2749 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c02);
2751 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c10);
2752 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2753 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c12);
2755 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c00);
2756 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c01);
2757 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c02);
2763 fGET10BIT(c00, VvvV.v[0].uw[i], 0)
2765 fGET10BIT(c01, VvvV.v[0].uw[i], 1)
2767 fGET10BIT(c02, VvvV.v[0].uw[i], 2)
2769 fGET10BIT(c10, VvvV.v[1].uw[i], 0)
2771 fGET10BIT(c11, VvvV.v[1].uw[i], 1)
2773 fGET10BIT(c12, VvvV.v[1].uw[i], 2)
2776 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c10);
2777 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c11);
2778 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c12);
2780 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c00);
2781 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2782 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c02);
2784 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c10);
2785 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c11);
2786 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c12);
2789 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c00);
2790 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c01);
2791 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c02);
2793 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(3,VuuV.v[1].uw[i]), c10);
2794 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c11);
2795 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c12);
2797 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[1].uw[i]), c00);
2798 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c01);
2799 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c02);
2802 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c10);
2803 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2804 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c12);
2806 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c00);
2807 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c01);
2808 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c02);
2810 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c10);
2811 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c11);
2812 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c12);
2815 VddV.v[1].w[i] = fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c00);
2816 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c01);
2817 VddV.v[1].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c02);
2819 VddV.v[0].w[i] = fMPY16US(fGETUBYTE(1,VuuV.v[1].uw[i]), c10);
2820 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(3,VuuV.v[0].uw[i]), c11);
2821 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(1,VuuV.v[0].uw[i]), c12);
2823 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[1].uw[i]), c00);
2824 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(2,VuuV.v[0].uw[i]), c01);
2825 VddV.v[0].w[i] += fMPY16US(fGETUBYTE(0,VuuV.v[0].uw[i]), c02);
2832 fHIDE(int i;)
2838 fVFOREACH(32, i) {
2840 EA = RtV+VvvV.v[j].uw[i];
2841 fVLOG_VTCM_HALFWORDQ_DV(EA,VvvV.v[j].uw[i],VwV,(2*i+j),QsV,i,j,MuV);
2849 fHIDE(int i;)
2856 fVFOREACH(32, i) {
2858 EA = RtV + fVALIGN(VvvV.v[j].uw[i],ALIGNMENT);
2859 fVLOG_VTCM_HALFWORD_INCREMENT_DV(EA,VvvV.v[j].uw[i],VwV,(2*i+j),i,j,ALIGNMENT,MuV);
2868 fHIDE(int i;)
2870 fVFOREACH(8, i) {
2871 acc += fGETQBIT(QvV,i);
2872 VdV.ub[i] = acc;
2877 fHIDE(int i;)
2879 fVFOREACH(16, i) {
2880 acc += fGETQBIT(QvV,i*2+0);
2881 acc += fGETQBIT(QvV,i*2+1);
2882 VdV.uh[i] = acc;
2887 fHIDE(int i;)
2889 fVFOREACH(32, i) {
2890 acc += fGETQBIT(QvV,i*4+0);
2891 acc += fGETQBIT(QvV,i*4+1);
2892 acc += fGETQBIT(QvV,i*4+2);
2893 acc += fGETQBIT(QvV,i*4+3);
2894 VdV.uw[i] = acc;
2907 int i; \
2910 for (i=0;i<vec_len>>COUNT;i++) { \
2911 fprintf(stdout,TYPE2 " ", VuV.TYPE[i]); \