📄 ddot.s
字号:
.align 32/* INCY == 1 */.L112: { .mmf (p16) LDFPD f32, f35 = [Y1], 2 * SIZE (p16) lfetch.nt1 [PREX], INCX16 (p18) FMA f8 = f34, f82, f8 } { .mmf (p16) LDFD f80 = [X1], INCX (p16) LDFD f86 = [X2], INCX (p18) FMA f9 = f37, f85, f9 } ;; { .mmf (p16) LDFPD f38, f41 = [Y1], 2 * SIZE (p16) lfetch.nt1 [PREY], INCY16 (p18) FMA f10 = f40, f88, f10 } { .mmf (p16) LDFD f83 = [X1], INCX3 (p16) LDFD f89 = [X2], INCX3 (p18) FMA f11 = f43, f91, f11 } ;; { .mmf (p16) LDFPD f44, f47 = [Y1], 2 * SIZE (p18) FMA f12 = f46, f94, f12 } { .mmf (p16) LDFD f92 = [X1], INCX (p16) LDFD f98 = [X2], INCX (p18) FMA f13 = f49, f97, f13 } ;; { .mmf (p16) LDFPD f50, f53 = [Y1], 2 * SIZE (p18) FMA f14 = f52, f100, f14 } { .mmf (p16) LDFD f95 = [X1], INCX3 (p16) LDFD f101 = [X2], INCX3 (p18) FMA f15 = f55, f103, f15 } ;; { .mmf (p16) LDFPD f56, f59 = [Y1], 2 * SIZE (p18) FMA f8 = f58, f106, f8 } { .mmf (p16) LDFD f104 = [X1], INCX (p16) LDFD f110 = [X2], INCX (p18) FMA f9 = f61, f109, f9 } ;; { .mmf (p16) LDFPD f62, f65 = [Y1], 2 * SIZE (p18) FMA f10 = f64, f112, f10 } { .mmf (p16) LDFD f107 = [X1], INCX3 (p16) LDFD f113 = [X2], INCX3 (p18) FMA f11 = f67, f115, f11 } ;; { .mmf (p16) LDFPD f68, f71 = [Y1], 2 * SIZE (p18) FMA f12 = f70, f118, f12 } { .mmf (p16) LDFD f116 = [X1], INCX (p16) LDFD f122 = [X2], INCX (p18) FMA f13 = f73, f121, f13 } ;; { .mmf (p16) LDFPD f74, f77 = [Y1], 2 * SIZE (p16) LDFD f119 = [X1], INCX3 (p18) FMA f14 = f76, f124, f14 } { .mfb (p16) LDFD f125 = [X2], INCX3 (p18) FMA f15 = f79, f127, f15 br.ctop.sptk.few .L112 } ;; .align 32.L115: { .mmi (p12) LDFPD f32, f33 = [Y1], 2 * SIZE mov XX = X1 tbit.z p0, p13 = N, 2 } { .mmb (p12) LDFD f34 = [X1], INCX (p12) LDFD f38 = [X2], INCX (p7) br.cond.dptk .L999 } ;; { .mmi (p12) LDFPD f36, f37 = [Y1], 2 * SIZE (p12) shladd XX = INCX, 3, XX tbit.z p0, p14 = N, 1 } { .mmi (p12) LDFD f35 = [X1], INCX3 (p12) LDFD f39 = [X2], INCX3 tbit.z p0, p15 = N, 0 } ;; { .mmi (p12) LDFPD f40, f41 = [Y1], 2 * SIZE (p13) shladd XX = INCX, 2, XX } { .mmi (p12) LDFD f42 = [X1], INCX (p12) LDFD f46 = [X2], INCX } ;; (p12) LDFPD f44, f45 = [Y1], 2 * SIZE (p12) LDFD f43 = [X1], INCX3 (p12) LDFD f47 = [X2], INCX3 (p14) shladd XX = INCX, 1, XX ;; (p13) LDFPD f48, f49 = [Y1], 2 * SIZE (p13) LDFD f50 = [X1], INCX (p13) LDFD f54 = [X2], INCX ;; (p13) LDFPD f52, f53 = [Y1], 2 * SIZE (p13) LDFD f51 = [X1], INCX3 (p13) LDFD f55 = [X2], INCX3 ;; (p14) LDFPD f56, f57 = [Y1], 2 * SIZE (p14) LDFD f58 = [X1], INCX (p15) LDFD f61 = [XX] ;; (p14) LDFD f59 = [X1] (p15) LDFD f60 = [Y1] ;; (p12) FMA f8 = f32, f34, f8 (p12) FMA f9 = f33, f35, f9 (p12) FMA f10 = f36, f38, f10 (p12) FMA f11 = f37, f39, f11 (p12) FMA f12 = f40, f42, f12 (p12) FMA f13 = f41, f43, f13 (p12) FMA f14 = f44, f46, f14 (p12) FMA f15 = f45, f47, f15 ;; (p13) FMA f8 = f48, f50, f8 (p13) FMA f9 = f49, f51, f9 (p13) FMA f10 = f52, f54, f10 (p13) FMA f11 = f53, f55, f11 (p14) FMA f12 = f56, f58, f12 (p14) FMA f13 = f57, f59, f13 (p15) FMA f14 = f60, f61, f14 br .L999 ;; .align 32.L120: { .mmi adds PREX = (PREFETCH_SIZE + 17) * SIZE, X1 adds PREY = (PREFETCH_SIZE + 19) * SIZE, X1 mov ar.lc = I } { .mfb cmp.eq p6 ,p0 = -1, I FMA f15 = f32, f80, f0 (p6) br.cond.dpnt .L125 } ;; .align 32.L122: { .mmf (p16) LDFPD f32, f35 = [Y1], 2 * SIZE (p16) lfetch.nt1 [PREX], INCX16 (p18) FMA f8 = f34, f82, f8 } { .mmf (p17) LDFD f105 = [X1], INCX (p17) LDFD f111 = [X2], INCX (p18) FMA f9 = f37, f85, f9 } ;; { .mmf (p16) LDFPD f38, f41 = [Y1], 2 * SIZE (p16) lfetch.nt1 [PREY], INCY16 (p18) FMA f10 = f40, f88, f10 } { .mmf (p17) LDFD f108 = [X1], INCX3 (p17) LDFD f114 = [X2], INCX3 (p18) FMA f11 = f43, f91, f11 } ;; { .mmf (p16) LDFPD f44, f47 = [Y1], 2 * SIZE (p18) FMA f12 = f46, f94, f12 } { .mmf (p17) LDFD f117 = [X1], INCX (p17) LDFD f123 = [X2], INCX (p18) FMA f13 = f49, f97, f13 } ;; { .mmf (p16) LDFPD f50, f53 = [Y1], 2 * SIZE (p18) FMA f14 = f52, f100, f14 } { .mmf (p17) LDFD f120 = [X1], INCX3 (p17) LDFD f126 = [X2], INCX3 (p18) FMA f15 = f55, f103, f15 } ;; { .mmf (p16) LDFPD f56, f59 = [Y1], 2 * SIZE (p18) FMA f8 = f58, f106, f8 } { .mmf (p16) LDFD f80 = [X1], INCX (p16) LDFD f86 = [X2], INCX (p18) FMA f9 = f61, f109, f9 } ;; { .mmf (p16) LDFPD f62, f65 = [Y1], 2 * SIZE (p18) FMA f10 = f64, f112, f10 } { .mmf (p16) LDFD f83 = [X1], INCX3 (p16) LDFD f89 = [X2], INCX3 (p18) FMA f11 = f67, f115, f11 } ;; { .mmf (p16) LDFPD f68, f71 = [Y1], 2 * SIZE (p18) FMA f12 = f70, f118, f12 } { .mmf (p16) LDFD f92 = [X1], INCX (p16) LDFD f98 = [X2], INCX (p18) FMA f13 = f73, f121, f13 } ;; { .mmf (p16) LDFPD f74, f77 = [Y1], 2 * SIZE (p16) LDFD f95 = [X1], INCX3 (p18) FMA f14 = f76, f124, f14 } { .mfb (p16) LDFD f101 = [X2], INCX3 (p18) FMA f15 = f79, f127, f15 br.ctop.sptk.few .L122 } ;; .align 32.L125: { .mmi (p12) LDFPD f32, f33 = [Y1], 2 * SIZE mov XX = X1 tbit.z p0, p13 = N, 2 } { .mmb (p12) LDFD f34 = [X1], INCX (p12) LDFD f38 = [X2], INCX (p7) br.cond.dptk .L999 } ;; { .mmi (p12) LDFPD f36, f37 = [Y1], 2 * SIZE (p12) shladd XX = INCX, 3, XX tbit.z p0, p14 = N, 1 } { .mmi (p12) LDFD f35 = [X1], INCX3 (p12) LDFD f39 = [X2], INCX3 tbit.z p0, p15 = N, 0 } ;; { .mmi (p12) LDFPD f40, f41 = [Y1], 2 * SIZE (p13) shladd XX = INCX, 2, XX } { .mmi (p12) LDFD f42 = [X1], INCX (p12) LDFD f46 = [X2], INCX } ;; (p12) LDFPD f44, f45 = [Y1], 2 * SIZE (p12) LDFD f43 = [X1], INCX3 (p12) LDFD f47 = [X2], INCX3 (p14) shladd XX = INCX, 1, XX ;; (p13) LDFPD f48, f49 = [Y1], 2 * SIZE (p13) LDFD f50 = [X1], INCX (p13) LDFD f54 = [X2], INCX ;; (p13) LDFPD f52, f53 = [Y1], 2 * SIZE (p13) LDFD f51 = [X1], INCX3 (p13) LDFD f55 = [X2], INCX3 ;; (p14) LDFPD f56, f57 = [Y1], 2 * SIZE (p14) LDFD f58 = [X1], INCX (p15) LDFD f61 = [XX] ;; (p14) LDFD f59 = [X1] (p15) LDFD f60 = [Y1] ;; (p12) FMA f8 = f32, f34, f8 (p12) FMA f9 = f33, f35, f9 (p12) FMA f10 = f36, f38, f10 (p12) FMA f11 = f37, f39, f11 (p12) FMA f12 = f40, f42, f12 (p12) FMA f13 = f41, f43, f13 (p12) FMA f14 = f44, f46, f14 (p12) FMA f15 = f45, f47, f15 ;; (p13) FMA f8 = f48, f50, f8 (p13) FMA f9 = f49, f51, f9 (p13) FMA f10 = f52, f54, f10 (p13) FMA f11 = f53, f55, f11 (p14) FMA f12 = f56, f58, f12 (p14) FMA f13 = f57, f59, f13 (p15) FMA f14 = f60, f61, f14 br .L999 ;; .align 32.L200: { .mfi shladd INCX3 = INCX, 1, INCX mov f12 = f0 mov pr.rot= 0 } { .mfi and J = 15, N mov f13 = f0 shr I = N, 4 } ;; { .mmf cmp.eq p16, p0 = r0, r0 shladd INCY3 = INCY, 1, INCY mov f14 = f0 } { .mmi shladd INCX16 = INCX, 4, r0 shladd INCY16 = INCY, 4, r0 tbit.z p0, p12 = N, 3 } ;; { .mmi cmp.eq p7, p0 = r0, J adds I = -1, I mov ar.ec= 3 } { .mmi shladd Y2 = INCY, 1, Y1 mov XX = X1 mov YY = Y1 } ;; { .mmi adds PREX = (PREFETCH_SIZE + 5) * SIZE, X1 adds PREY = (PREFETCH_SIZE + 3) * SIZE, Y1 mov ar.lc = I } { .mfb cmp.eq p6 ,p0 = -1, I mov f15 = f0 (p6) br.cond.dpnt .L215 } ;; .align 32/* INCY == 1 */.L212: { .mmf (p16) lfetch.nt1 [PREX], INCX16 (p16) lfetch.nt1 [PREY], INCY16 (p18) FMA f8 = f34, f82, f8 } { .mmf (p16) LDFD f32 = [Y1], INCY (p16) LDFD f38 = [Y2], INCY (p18) FMA f9 = f37, f85, f9 } ;; { .mmf (p16) LDFD f80 = [X1], INCX (p16) LDFD f86 = [X2], INCX (p18) FMA f10 = f40, f88, f10 } { .mmf (p16) LDFD f35 = [Y1], INCY3 (p16) LDFD f41 = [Y2], INCY3 (p18) FMA f11 = f43, f91, f11 } ;; { .mmf (p16) LDFD f83 = [X1], INCX3 (p16) LDFD f89 = [X2], INCX3 (p18) FMA f12 = f46, f94, f12 } { .mmf (p16) LDFD f44 = [Y1], INCY (p16) LDFD f50 = [Y2], INCY (p18) FMA f13 = f49, f97, f13 } ;; { .mmf (p16) LDFD f92 = [X1], INCX (p16) LDFD f98 = [X2], INCX (p18) FMA f14 = f52, f100, f14 } { .mmf (p16) LDFD f47 = [Y1], INCY3 (p16) LDFD f53 = [Y2], INCY3 (p18) FMA f15 = f55, f103, f15 } ;; { .mmf (p16) LDFD f95 = [X1], INCX3 (p16) LDFD f101 = [X2], INCX3 (p18) FMA f8 = f58, f106, f8 } { .mmf (p16) LDFD f56 = [Y1], INCY (p16) LDFD f62 = [Y2], INCY (p18) FMA f9 = f61, f109, f9 } ;; { .mmf (p16) LDFD f104 = [X1], INCX (p16) LDFD f110 = [X2], INCX (p18) FMA f10 = f64, f112, f10 } { .mmf (p16) LDFD f59 = [Y1], INCY3 (p16) LDFD f65 = [Y2], INCY3 (p18) FMA f11 = f67, f115, f11 } ;; { .mmf (p16) LDFD f107 = [X1], INCX3 (p16) LDFD f113 = [X2], INCX3 (p18) FMA f12 = f70, f118, f12 } { .mmf (p16) LDFD f68 = [Y1], INCY (p16) LDFD f74 = [Y2], INCY (p18) FMA f13 = f73, f121, f13 } ;; { .mmf (p16) LDFD f116 = [X1], INCX (p16) LDFD f122 = [X2], INCX (p18) FMA f14 = f76, f124, f14 } { .mmf (p16) LDFD f71 = [Y1], INCY3 (p16) LDFD f77 = [Y2], INCY3 (p18) FMA f15 = f79, f127, f15 } ;; { .mmi (p16) LDFD f119 = [X1], INCX3 (p16) LDFD f125 = [X2], INCX3 } { .mmb (p16) add XX = INCX16, XX (p16) add YY = INCY16, YY br.ctop.sptk.few .L212 } ;; .align 32.L215: { .mmi (p12) LDFD f34 = [X1], INCX (p12) LDFD f38 = [X2], INCX tbit.z p0, p13 = N, 2 } { .mmb (p12) LDFD f32 = [Y1], INCY (p12) LDFD f36 = [Y2], INCY (p7) br.cond.dptk .L999 } ;; { .mmi (p12) LDFD f35 = [X1], INCX3 (p12) LDFD f39 = [X2], INCX3 tbit.z p0, p14 = N, 1 } { .mmi (p12) LDFD f33 = [Y1], INCY3 (p12) LDFD f37 = [Y2], INCY3 tbit.z p0, p15 = N, 0 } ;; { .mmi (p12) LDFD f42 = [X1], INCX (p12) LDFD f46 = [X2], INCX (p12) shladd XX = INCX, 3, XX } { .mmi (p12) LDFD f40 = [Y1], INCY (p12) LDFD f44 = [Y2], INCY (p12) shladd YY = INCY, 3, YY } ;; { .mmi (p12) LDFD f43 = [X1], INCX3 (p12) LDFD f47 = [X2], INCX3 (p13) shladd XX = INCX, 2, XX } { .mmi (p12) LDFD f41 = [Y1], INCY3 (p12) LDFD f45 = [Y2], INCY3 (p13) shladd YY = INCY, 2, YY } ;; (p13) LDFD f50 = [X1], INCX (p13) LDFD f54 = [X2], INCX (p14) shladd XX = INCX, 1, XX (p13) LDFD f48 = [Y1], INCY (p13) LDFD f52 = [Y2], INCY (p14) shladd YY = INCY, 1, YY ;; (p13) LDFD f51 = [X1], INCX3 (p13) LDFD f55 = [X2] (p13) LDFD f49 = [Y1], INCY3 (p13) LDFD f53 = [Y2] ;; (p14) LDFD f58 = [X1], INCX (p15) LDFD f61 = [XX] (p14) LDFD f56 = [Y1], INCY (p15) LDFD f60 = [YY] ;; (p14) LDFD f59 = [X1] (p14) LDFD f57 = [Y1] ;; ;; ;; (p12) FMA f8 = f32, f34, f8 (p12) FMA f9 = f33, f35, f9 (p12) FMA f10 = f36, f38, f10 (p12) FMA f11 = f37, f39, f11 (p12) FMA f12 = f40, f42, f12 (p12) FMA f13 = f41, f43, f13 (p12) FMA f14 = f44, f46, f14 (p12) FMA f15 = f45, f47, f15 ;; (p13) FMA f8 = f48, f50, f8 (p13) FMA f9 = f49, f51, f9 (p13) FMA f10 = f52, f54, f10 (p13) FMA f11 = f53, f55, f11 (p14) FMA f12 = f56, f58, f12 (p14) FMA f13 = f57, f59, f13 (p15) FMA f14 = f60, f61, f14 ;; .align 32.L999: FADD f8 = f8, f9 FADD f10 = f10, f11 FADD f12 = f12, f13 FADD f14 = f14, f15 ;; FADD f8 = f8, f10 FADD f12 = f12, f14 mov ar.lc = ARLC ;; FADD f8 = f8, f12 mov pr = PR, -65474 br.ret.sptk.many b0 EPILOGUE
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -