diff options
Diffstat (limited to 'media/ffvpx/libavcodec/x86/h264_intrapred.asm')
| -rw-r--r-- | media/ffvpx/libavcodec/x86/h264_intrapred.asm | 31 |
1 files changed, 13 insertions, 18 deletions
diff --git a/media/ffvpx/libavcodec/x86/h264_intrapred.asm b/media/ffvpx/libavcodec/x86/h264_intrapred.asm index 8a38ba2bb5..a8a630dbe6 100644 --- a/media/ffvpx/libavcodec/x86/h264_intrapred.asm +++ b/media/ffvpx/libavcodec/x86/h264_intrapred.asm @@ -86,8 +86,6 @@ cglobal pred16x16_horizontal_8, 2,3 punpcklbw m1, m1 SPLATW m0, m0, 3 SPLATW m1, m1, 3 - mova [r0+r1*0+8], m0 - mova [r0+r1*1+8], m1 %endif mova [r0+r1*0], m0 @@ -98,7 +96,7 @@ cglobal pred16x16_horizontal_8, 2,3 RET %endmacro -INIT_MMX mmxext +INIT_XMM sse2 PRED16x16_H INIT_XMM ssse3 PRED16x16_H @@ -568,17 +566,17 @@ H264_PRED8x8_PLANE ; void ff_pred8x8_vertical_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- -INIT_MMX mmx +INIT_XMM sse2 cglobal pred8x8_vertical_8, 2,2 sub r0, r1 - movq mm0, [r0] + movq m0, [r0] %rep 3 - movq [r0+r1*1], mm0 - movq [r0+r1*2], mm0 + movq [r0+r1*1], m0 + movq [r0+r1*2], m0 lea r0, [r0+r1*2] %endrep - movq [r0+r1*1], mm0 - movq [r0+r1*2], mm0 + movq [r0+r1*1], m0 + movq [r0+r1*2], m0 RET ;----------------------------------------------------------------------------- @@ -1313,10 +1311,7 @@ PRED8x8L_DOWN_RIGHT ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL_RIGHT 0 -cglobal pred8x8l_vertical_right_8, 4,5,7 - ; manually spill XMM registers for Win64 because - ; the code here is initialized with INIT_MMX - WIN64_SPILL_XMM 7 +cglobal pred8x8l_vertical_right_8, 4,5,6 sub r0, r3 lea r4, [r0+r3*2] movq mm0, [r0+r3*1-8] @@ -1386,7 +1381,6 @@ cglobal pred8x8l_vertical_right_8, 4,5,7 movq2dq xmm4, mm6 pslldq xmm4, 8 por xmm0, xmm4 - movdqa xmm6, [pw_ff00] movdqa xmm1, xmm0 lea r2, [r1+r3*2] movdqa xmm2, xmm0 @@ -1396,15 +1390,16 @@ cglobal pred8x8l_vertical_right_8, 4,5,7 pavgb xmm2, xmm0 INIT_XMM cpuname PRED4x4_LOWPASS xmm4, xmm3, xmm1, xmm0, xmm5 - pandn xmm6, xmm4 + movdqa xmm0, [pw_ff00] + pandn xmm0, xmm4 movdqa xmm5, xmm4 psrlw xmm4, 8 - packuswb xmm6, xmm4 - movhlps xmm4, xmm6 + packuswb xmm0, xmm4 + movhlps xmm4, xmm0 movhps [r0+r3*2], xmm5 movhps [r0+r3*1], xmm2 psrldq xmm5, 4 - movss xmm5, xmm6 + movss xmm5, xmm0 psrldq xmm2, 4 movss xmm2, xmm4 lea r0, [r2+r3*2] |