ffmpeg | branch: master | Clément Bœsch <u...@pkh.me> | Thu Mar 23 11:58:01 2017 +0100| [4bb4fa28e3741663e0e76f2e2a628cfdcca6765c] | committer: Clément Bœsch
Merge commit '5801f9ed245ca5ebb57b0b5183de7a24aaece133' * commit '5801f9ed245ca5ebb57b0b5183de7a24aaece133': h264_intrapred: x86: Update comments left behind in 95c89da36ebeeb96b7146c0d70f46c582397da7f Merged-by: Clément Bœsch <u...@pkh.me> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4bb4fa28e3741663e0e76f2e2a628cfdcca6765c --- libavcodec/x86/h264_intrapred.asm | 65 +++++++++++++++------------- libavcodec/x86/h264_intrapred_10bit.asm | 77 ++++++++++++++++++--------------- 2 files changed, 76 insertions(+), 66 deletions(-) diff --git a/libavcodec/x86/h264_intrapred.asm b/libavcodec/x86/h264_intrapred.asm index 0f3b462..f3aa317 100644 --- a/libavcodec/x86/h264_intrapred.asm +++ b/libavcodec/x86/h264_intrapred.asm @@ -49,7 +49,7 @@ cextern pw_17 cextern pw_32 ;----------------------------------------------------------------------------- -; void ff_pred16x16_vertical_8(uint8_t *src, int stride) +; void ff_pred16x16_vertical_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmx @@ -85,7 +85,7 @@ cglobal pred16x16_vertical_8, 2,3 REP_RET ;----------------------------------------------------------------------------- -; void ff_pred16x16_horizontal_8(uint8_t *src, int stride) +; void ff_pred16x16_horizontal_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_H 0 @@ -126,7 +126,7 @@ INIT_XMM ssse3 PRED16x16_H ;----------------------------------------------------------------------------- -; void ff_pred16x16_dc_8(uint8_t *src, int stride) +; void ff_pred16x16_dc_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_DC 0 @@ -188,7 +188,7 @@ INIT_XMM ssse3 PRED16x16_DC ;----------------------------------------------------------------------------- -; void ff_pred16x16_tm_vp8_8(uint8_t *src, int stride) +; void ff_pred16x16_tm_vp8_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_TM 0 @@ -306,7 +306,7 @@ cglobal pred16x16_tm_vp8_8, 2, 4, 5, dst, stride, stride3, iteration %endif ;----------------------------------------------------------------------------- -; void ff_pred16x16_plane_*_8(uint8_t *src, int stride) +; void ff_pred16x16_plane_*_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro H264_PRED16x16_PLANE 1 @@ -587,7 +587,7 @@ H264_PRED16x16_PLANE rv40 H264_PRED16x16_PLANE svq3 ;----------------------------------------------------------------------------- -; void ff_pred8x8_plane_8(uint8_t *src, int stride) +; void ff_pred8x8_plane_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro H264_PRED8x8_PLANE 0 @@ -761,7 +761,7 @@ INIT_XMM ssse3 H264_PRED8x8_PLANE ;----------------------------------------------------------------------------- -; void ff_pred8x8_vertical_8(uint8_t *src, int stride) +; void ff_pred8x8_vertical_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmx @@ -778,7 +778,7 @@ cglobal pred8x8_vertical_8, 2,2 RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_horizontal_8(uint8_t *src, int stride) +; void ff_pred8x8_horizontal_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8_H 0 @@ -806,7 +806,7 @@ INIT_MMX ssse3 PRED8x8_H ;----------------------------------------------------------------------------- -; void ff_pred8x8_top_dc_8_mmxext(uint8_t *src, int stride) +; void ff_pred8x8_top_dc_8_mmxext(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext cglobal pred8x8_top_dc_8, 2,5 @@ -840,7 +840,7 @@ cglobal pred8x8_top_dc_8, 2,5 RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_dc_8_mmxext(uint8_t *src, int stride) +; void ff_pred8x8_dc_8_mmxext(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -901,7 +901,7 @@ cglobal pred8x8_dc_8, 2,5 RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_dc_rv40_8(uint8_t *src, int stride) +; void ff_pred8x8_dc_rv40_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -938,7 +938,7 @@ cglobal pred8x8_dc_rv40_8, 2,7 REP_RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_tm_vp8_8(uint8_t *src, int stride) +; void ff_pred8x8_tm_vp8_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8_TM 0 @@ -1051,7 +1051,7 @@ cglobal pred8x8_tm_vp8_8, 2,3,6 ;----------------------------------------------------------------------------- ; void ff_pred8x8l_top_dc_8(uint8_t *src, int has_topleft, int has_topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_TOP_DC 0 cglobal pred8x8l_top_dc_8, 4,4 @@ -1107,7 +1107,7 @@ PRED8x8L_TOP_DC ;----------------------------------------------------------------------------- ; void ff_pred8x8l_dc_8(uint8_t *src, int has_topleft, int has_topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_DC 0 @@ -1211,7 +1211,7 @@ PRED8x8L_DC ;----------------------------------------------------------------------------- ; void ff_pred8x8l_horizontal_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL 0 @@ -1283,7 +1283,7 @@ PRED8x8L_HORIZONTAL ;----------------------------------------------------------------------------- ; void ff_pred8x8l_vertical_8(uint8_t *src, int has_topleft, int has_topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL 0 @@ -1334,7 +1334,7 @@ PRED8x8L_VERTICAL ;----------------------------------------------------------------------------- ; void ff_pred8x8l_down_left_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -1535,7 +1535,7 @@ PRED8x8L_DOWN_LEFT ;----------------------------------------------------------------------------- ; void ff_pred8x8l_down_right_8_mmxext(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -1787,7 +1787,7 @@ PRED8x8L_DOWN_RIGHT ;----------------------------------------------------------------------------- ; void ff_pred8x8l_vertical_right_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -2015,7 +2015,7 @@ PRED8x8L_VERTICAL_RIGHT ;----------------------------------------------------------------------------- ; void ff_pred8x8l_vertical_left_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL_LEFT 0 @@ -2105,7 +2105,7 @@ PRED8x8L_VERTICAL_LEFT ;----------------------------------------------------------------------------- ; void ff_pred8x8l_horizontal_up_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL_UP 0 @@ -2193,7 +2193,7 @@ PRED8x8L_HORIZONTAL_UP ;----------------------------------------------------------------------------- ; void ff_pred8x8l_horizontal_down_8(uint8_t *src, int has_topleft, -; int has_topright, int stride) +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -2441,7 +2441,8 @@ INIT_MMX ssse3 PRED8x8L_HORIZONTAL_DOWN ;------------------------------------------------------------------------------- -; void ff_pred4x4_dc_8_mmxext(uint8_t *src, const uint8_t *topright, int stride) +; void ff_pred4x4_dc_8_mmxext(uint8_t *src, const uint8_t *topright, +; ptrdiff_t stride) ;------------------------------------------------------------------------------- INIT_MMX mmxext @@ -2472,7 +2473,7 @@ cglobal pred4x4_dc_8, 3,5 ;----------------------------------------------------------------------------- ; void ff_pred4x4_tm_vp8_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED4x4_TM 0 @@ -2551,7 +2552,7 @@ cglobal pred4x4_tm_vp8_8, 3,3 ;----------------------------------------------------------------------------- ; void ff_pred4x4_vertical_vp8_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -2572,7 +2573,7 @@ cglobal pred4x4_vertical_vp8_8, 3,3 ;----------------------------------------------------------------------------- ; void ff_pred4x4_down_left_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext cglobal pred4x4_down_left_8, 3,3 @@ -2599,7 +2600,7 @@ cglobal pred4x4_down_left_8, 3,3 ;------------------------------------------------------------------------------ ; void ff_pred4x4_vertical_left_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;------------------------------------------------------------------------------ INIT_MMX mmxext @@ -2625,7 +2626,7 @@ cglobal pred4x4_vertical_left_8, 3,3 ;------------------------------------------------------------------------------ ; void ff_pred4x4_horizontal_up_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;------------------------------------------------------------------------------ INIT_MMX mmxext @@ -2659,7 +2660,8 @@ cglobal pred4x4_horizontal_up_8, 3,3 ;------------------------------------------------------------------------------ ; void ff_pred4x4_horizontal_down_8_mmxext(uint8_t *src, -; const uint8_t *topright, int stride) +; const uint8_t *topright, +; ptrdiff_t stride) ;------------------------------------------------------------------------------ INIT_MMX mmxext @@ -2695,7 +2697,8 @@ cglobal pred4x4_horizontal_down_8, 3,3 ;----------------------------------------------------------------------------- ; void ff_pred4x4_vertical_right_8_mmxext(uint8_t *src, -; const uint8_t *topright, int stride) +; const uint8_t *topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -2726,7 +2729,7 @@ cglobal pred4x4_vertical_right_8, 3,3 ;----------------------------------------------------------------------------- ; void ff_pred4x4_down_right_8_mmxext(uint8_t *src, const uint8_t *topright, -; int stride) +; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext diff --git a/libavcodec/x86/h264_intrapred_10bit.asm b/libavcodec/x86/h264_intrapred_10bit.asm index 9e40cfe..629e0a7 100644 --- a/libavcodec/x86/h264_intrapred_10bit.asm +++ b/libavcodec/x86/h264_intrapred_10bit.asm @@ -51,7 +51,8 @@ SECTION .text %endmacro ;----------------------------------------------------------------------------- -; void ff_pred4x4_down_right(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_down_right_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED4x4_DR 0 cglobal pred4x4_down_right_10, 3, 3 @@ -89,7 +90,8 @@ PRED4x4_DR %endif ;------------------------------------------------------------------------------ -; void ff_pred4x4_vertical_right(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_vertical_right_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;------------------------------------------------------------------------------ %macro PRED4x4_VR 0 cglobal pred4x4_vertical_right_10, 3, 3, 6 @@ -128,7 +130,8 @@ PRED4x4_VR %endif ;------------------------------------------------------------------------------- -; void ff_pred4x4_horizontal_down(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_horizontal_down_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;------------------------------------------------------------------------------- %macro PRED4x4_HD 0 cglobal pred4x4_horizontal_down_10, 3, 3 @@ -170,7 +173,7 @@ PRED4x4_HD %endif ;----------------------------------------------------------------------------- -; void ff_pred4x4_dc(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_dc_10(pixel *src, const pixel *topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext @@ -195,7 +198,8 @@ cglobal pred4x4_dc_10, 3, 3 RET ;----------------------------------------------------------------------------- -; void ff_pred4x4_down_left(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_down_left_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED4x4_DL 0 cglobal pred4x4_down_left_10, 3, 3 @@ -225,7 +229,8 @@ PRED4x4_DL %endif ;----------------------------------------------------------------------------- -; void ff_pred4x4_vertical_left(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_vertical_left_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED4x4_VL 0 cglobal pred4x4_vertical_left_10, 3, 3 @@ -254,7 +259,8 @@ PRED4x4_VL %endif ;----------------------------------------------------------------------------- -; void ff_pred4x4_horizontal_up(pixel *src, const pixel *topright, int stride) +; void ff_pred4x4_horizontal_up_10(pixel *src, const pixel *topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext cglobal pred4x4_horizontal_up_10, 3, 3 @@ -288,7 +294,7 @@ cglobal pred4x4_horizontal_up_10, 3, 3 ;----------------------------------------------------------------------------- -; void ff_pred8x8_vertical(pixel *src, int stride) +; void ff_pred8x8_vertical_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_XMM sse2 cglobal pred8x8_vertical_10, 2, 2 @@ -304,7 +310,7 @@ cglobal pred8x8_vertical_10, 2, 2 RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_horizontal(pixel *src, int stride) +; void ff_pred8x8_horizontal_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_XMM sse2 cglobal pred8x8_horizontal_10, 2, 3 @@ -324,7 +330,7 @@ cglobal pred8x8_horizontal_10, 2, 3 REP_RET ;----------------------------------------------------------------------------- -; void ff_predict_8x8_dc(pixel *src, int stride) +; void ff_predict_8x8_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro MOV8 2-3 ; sort of a hack, but it works @@ -411,7 +417,7 @@ INIT_XMM sse2 PRED8x8_DC pshuflw ;----------------------------------------------------------------------------- -; void ff_pred8x8_top_dc(pixel *src, int stride) +; void ff_pred8x8_top_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_XMM sse2 cglobal pred8x8_top_dc_10, 2, 4 @@ -438,7 +444,7 @@ cglobal pred8x8_top_dc_10, 2, 4 RET ;----------------------------------------------------------------------------- -; void ff_pred8x8_plane(pixel *src, int stride) +; void ff_pred8x8_plane_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_XMM sse2 cglobal pred8x8_plane_10, 2, 7, 7 @@ -501,8 +507,8 @@ cglobal pred8x8_plane_10, 2, 7, 7 ;----------------------------------------------------------------------------- -; void ff_pred8x8l_128_dc(pixel *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_128_dc_10(pixel *src, int has_topleft, int has_topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_128_DC 0 cglobal pred8x8l_128_dc_10, 4, 4 @@ -526,8 +532,8 @@ INIT_XMM sse2 PRED8x8L_128_DC ;----------------------------------------------------------------------------- -; void ff_pred8x8l_top_dc(pixel *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_top_dc_10(pixel *src, int has_topleft, int has_topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_TOP_DC 0 cglobal pred8x8l_top_dc_10, 4, 4, 6 @@ -566,7 +572,8 @@ PRED8x8L_TOP_DC %endif ;------------------------------------------------------------------------------- -; void ff_pred8x8l_dc(pixel *src, int has_topleft, int has_topright, int stride) +; void ff_pred8x8l_dc_10(pixel *src, int has_topleft, int has_topright, +; ptrdiff_t stride) ;------------------------------------------------------------------------------- ;TODO: see if scalar is faster %macro PRED8x8L_DC 0 @@ -625,8 +632,8 @@ PRED8x8L_DC %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_vertical(pixel *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_vertical_10(pixel *src, int has_topleft, int has_topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL 0 cglobal pred8x8l_vertical_10, 4, 4, 6 @@ -661,8 +668,8 @@ PRED8x8L_VERTICAL %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_horizontal(uint8_t *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_horizontal_10(uint8_t *src, int has_topleft, +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL 0 cglobal pred8x8l_horizontal_10, 4, 4, 5 @@ -718,8 +725,8 @@ PRED8x8L_HORIZONTAL %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_down_left(pixel *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_down_left_10(pixel *src, int has_topleft, int has_topright, +; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_DOWN_LEFT 0 cglobal pred8x8l_down_left_10, 4, 4, 7 @@ -787,8 +794,8 @@ PRED8x8L_DOWN_LEFT %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_down_right(pixel *src, int has_topleft, int has_topright, -; int stride) +; void ff_pred8x8l_down_right_10(pixel *src, int has_topleft, +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_DOWN_RIGHT 0 ; standard forbids this when has_topleft is false @@ -862,8 +869,8 @@ PRED8x8L_DOWN_RIGHT %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_vertical_right(pixel *src, int has_topleft, -; int has_topright, int stride) +; void ff_pred8x8l_vertical_right_10(pixel *src, int has_topleft, +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL_RIGHT 0 ; likewise with 8x8l_down_right @@ -933,8 +940,8 @@ PRED8x8L_VERTICAL_RIGHT %endif ;----------------------------------------------------------------------------- -; void ff_pred8x8l_horizontal_up(pixel *src, int has_topleft, -; int has_topright, int stride) +; void ff_pred8x8l_horizontal_up_10(pixel *src, int has_topleft, +; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL_UP 0 cglobal pred8x8l_horizontal_up_10, 4, 4, 6 @@ -996,7 +1003,7 @@ PRED8x8L_HORIZONTAL_UP ;----------------------------------------------------------------------------- -; void ff_pred16x16_vertical(pixel *src, int stride) +; void ff_pred16x16_vertical_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro MOV16 3-5 mova [%1+ 0], %2 @@ -1032,7 +1039,7 @@ INIT_XMM sse2 PRED16x16_VERTICAL ;----------------------------------------------------------------------------- -; void ff_pred16x16_horizontal(pixel *src, int stride) +; void ff_pred16x16_horizontal_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_HORIZONTAL 0 cglobal pred16x16_horizontal_10, 2, 3 @@ -1056,7 +1063,7 @@ INIT_XMM sse2 PRED16x16_HORIZONTAL ;----------------------------------------------------------------------------- -; void ff_pred16x16_dc(pixel *src, int stride) +; void ff_pred16x16_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_DC 0 cglobal pred16x16_dc_10, 2, 6 @@ -1102,7 +1109,7 @@ INIT_XMM sse2 PRED16x16_DC ;----------------------------------------------------------------------------- -; void ff_pred16x16_top_dc(pixel *src, int stride) +; void ff_pred16x16_top_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_TOP_DC 0 cglobal pred16x16_top_dc_10, 2, 3 @@ -1134,7 +1141,7 @@ INIT_XMM sse2 PRED16x16_TOP_DC ;----------------------------------------------------------------------------- -; void ff_pred16x16_left_dc(pixel *src, int stride) +; void ff_pred16x16_left_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_LEFT_DC 0 cglobal pred16x16_left_dc_10, 2, 6 @@ -1171,7 +1178,7 @@ INIT_XMM sse2 PRED16x16_LEFT_DC ;----------------------------------------------------------------------------- -; void ff_pred16x16_128_dc(pixel *src, int stride) +; void ff_pred16x16_128_dc_10(pixel *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED16x16_128_DC 0 cglobal pred16x16_128_dc_10, 2,3 ====================================================================== diff --cc libavcodec/x86/h264_intrapred.asm index 0f3b462,1ea97fa..f3aa317 --- a/libavcodec/x86/h264_intrapred.asm +++ b/libavcodec/x86/h264_intrapred.asm @@@ -268,45 -268,8 +268,45 @@@ cglobal pred16x16_tm_vp8_8, 2,6, jg .loop REP_RET +%if HAVE_AVX2_EXTERNAL +INIT_YMM avx2 +cglobal pred16x16_tm_vp8_8, 2, 4, 5, dst, stride, stride3, iteration + sub dstq, strideq + pmovzxbw m0, [dstq] + vpbroadcastb xm1, [r0-1] + pmovzxbw m1, xm1 + psubw m0, m1 + mov iterationd, 4 + lea stride3q, [strideq*3] +.loop: + vpbroadcastb xm1, [dstq+strideq*1-1] + vpbroadcastb xm2, [dstq+strideq*2-1] + vpbroadcastb xm3, [dstq+stride3q-1] + vpbroadcastb xm4, [dstq+strideq*4-1] + pmovzxbw m1, xm1 + pmovzxbw m2, xm2 + pmovzxbw m3, xm3 + pmovzxbw m4, xm4 + paddw m1, m0 + paddw m2, m0 + paddw m3, m0 + paddw m4, m0 + vpackuswb m1, m1, m2 + vpackuswb m3, m3, m4 + vpermq m1, m1, q3120 + vpermq m3, m3, q3120 + movdqa [dstq+strideq*1], xm1 + vextracti128 [dstq+strideq*2], m1, 1 + movdqa [dstq+stride3q*1], xm3 + vextracti128 [dstq+strideq*4], m3, 1 + lea dstq, [dstq+strideq*4] + dec iterationd + jg .loop + REP_RET +%endif + ;----------------------------------------------------------------------------- - ; void ff_pred16x16_plane_*_8(uint8_t *src, int stride) + ; void ff_pred16x16_plane_*_8(uint8_t *src, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro H264_PRED16x16_PLANE 1 diff --cc libavcodec/x86/h264_intrapred_10bit.asm index 9e40cfe,7ba9828..629e0a7 --- a/libavcodec/x86/h264_intrapred_10bit.asm +++ b/libavcodec/x86/h264_intrapred_10bit.asm @@@ -83,13 -83,12 +84,14 @@@ INIT_XMM sse PRED4x4_DR INIT_XMM ssse3 PRED4x4_DR +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED4x4_DR +%endif ;------------------------------------------------------------------------------ - ; void ff_pred4x4_vertical_right(pixel *src, const pixel *topright, int stride) + ; void ff_pred4x4_vertical_right_10(pixel *src, const pixel *topright, + ; ptrdiff_t stride) ;------------------------------------------------------------------------------ %macro PRED4x4_VR 0 cglobal pred4x4_vertical_right_10, 3, 3, 6 @@@ -122,13 -121,12 +124,14 @@@ INIT_XMM sse PRED4x4_VR INIT_XMM ssse3 PRED4x4_VR +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED4x4_VR +%endif ;------------------------------------------------------------------------------- - ; void ff_pred4x4_horizontal_down(pixel *src, const pixel *topright, int stride) + ; void ff_pred4x4_horizontal_down_10(pixel *src, const pixel *topright, + ; ptrdiff_t stride) ;------------------------------------------------------------------------------- %macro PRED4x4_HD 0 cglobal pred4x4_horizontal_down_10, 3, 3 @@@ -164,14 -162,28 +167,14 @@@ INIT_XMM sse PRED4x4_HD INIT_XMM ssse3 PRED4x4_HD +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED4x4_HD +%endif ;----------------------------------------------------------------------------- - ; void ff_pred4x4_dc(pixel *src, const pixel *topright, int stride) + ; void ff_pred4x4_dc_10(pixel *src, const pixel *topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- -%macro HADDD 2 ; sum junk -%if mmsize == 16 - movhlps %2, %1 - paddd %1, %2 - pshuflw %2, %1, 0xE - paddd %1, %2 -%else - pshufw %2, %1, 0xE - paddd %1, %2 -%endif -%endmacro - -%macro HADDW 2 - pmaddwd %1, [pw_1] - HADDD %1, %2 -%endmacro INIT_MMX mmxext cglobal pred4x4_dc_10, 3, 3 @@@ -219,13 -232,12 +223,14 @@@ cglobal pred4x4_down_left_10, 3, INIT_XMM sse2 PRED4x4_DL +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED4x4_DL +%endif ;----------------------------------------------------------------------------- - ; void ff_pred4x4_vertical_left(pixel *src, const pixel *topright, int stride) + ; void ff_pred4x4_vertical_left_10(pixel *src, const pixel *topright, + ; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED4x4_VL 0 cglobal pred4x4_vertical_left_10, 3, 3 @@@ -248,13 -260,12 +253,14 @@@ INIT_XMM sse2 PRED4x4_VL +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED4x4_VL +%endif ;----------------------------------------------------------------------------- - ; void ff_pred4x4_horizontal_up(pixel *src, const pixel *topright, int stride) + ; void ff_pred4x4_horizontal_up_10(pixel *src, const pixel *topright, + ; ptrdiff_t stride) ;----------------------------------------------------------------------------- INIT_MMX mmxext cglobal pred4x4_horizontal_up_10, 3, 3 @@@ -560,13 -571,12 +566,14 @@@ cglobal pred8x8l_top_dc_10, 4, 4, INIT_XMM sse2 PRED8x8L_TOP_DC +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_TOP_DC +%endif ;------------------------------------------------------------------------------- - ; void ff_pred8x8l_dc(pixel *src, int has_topleft, int has_topright, int stride) + ; void ff_pred8x8l_dc_10(pixel *src, int has_topleft, int has_topright, + ; ptrdiff_t stride) ;------------------------------------------------------------------------------- ;TODO: see if scalar is faster %macro PRED8x8L_DC 0 @@@ -619,14 -629,12 +626,14 @@@ cglobal pred8x8l_dc_10, 4, 6, INIT_XMM sse2 PRED8x8L_DC +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_DC +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_vertical(pixel *src, int has_topleft, int has_topright, - ; int stride) + ; void ff_pred8x8l_vertical_10(pixel *src, int has_topleft, int has_topright, + ; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL 0 cglobal pred8x8l_vertical_10, 4, 4, 6 @@@ -655,14 -663,12 +662,14 @@@ INIT_XMM sse2 PRED8x8L_VERTICAL +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_VERTICAL +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_horizontal(uint8_t *src, int has_topleft, int has_topright, - ; int stride) + ; void ff_pred8x8l_horizontal_10(uint8_t *src, int has_topleft, + ; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL 0 cglobal pred8x8l_horizontal_10, 4, 4, 5 @@@ -712,14 -718,12 +719,14 @@@ INIT_XMM sse PRED8x8L_HORIZONTAL INIT_XMM ssse3 PRED8x8L_HORIZONTAL +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_HORIZONTAL +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_down_left(pixel *src, int has_topleft, int has_topright, - ; int stride) + ; void ff_pred8x8l_down_left_10(pixel *src, int has_topleft, int has_topright, + ; ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_DOWN_LEFT 0 cglobal pred8x8l_down_left_10, 4, 4, 7 @@@ -781,14 -785,12 +788,14 @@@ INIT_XMM sse PRED8x8L_DOWN_LEFT INIT_XMM ssse3 PRED8x8L_DOWN_LEFT +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_DOWN_LEFT +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_down_right(pixel *src, int has_topleft, int has_topright, - ; int stride) + ; void ff_pred8x8l_down_right_10(pixel *src, int has_topleft, + ; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_DOWN_RIGHT 0 ; standard forbids this when has_topleft is false @@@ -856,14 -858,12 +863,14 @@@ INIT_XMM sse PRED8x8L_DOWN_RIGHT INIT_XMM ssse3 PRED8x8L_DOWN_RIGHT +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_DOWN_RIGHT +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_vertical_right(pixel *src, int has_topleft, - ; int has_topright, int stride) + ; void ff_pred8x8l_vertical_right_10(pixel *src, int has_topleft, + ; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_VERTICAL_RIGHT 0 ; likewise with 8x8l_down_right @@@ -927,14 -927,12 +934,14 @@@ INIT_XMM sse PRED8x8L_VERTICAL_RIGHT INIT_XMM ssse3 PRED8x8L_VERTICAL_RIGHT +%if HAVE_AVX_EXTERNAL INIT_XMM avx PRED8x8L_VERTICAL_RIGHT +%endif ;----------------------------------------------------------------------------- - ; void ff_pred8x8l_horizontal_up(pixel *src, int has_topleft, - ; int has_topright, int stride) + ; void ff_pred8x8l_horizontal_up_10(pixel *src, int has_topleft, + ; int has_topright, ptrdiff_t stride) ;----------------------------------------------------------------------------- %macro PRED8x8L_HORIZONTAL_UP 0 cglobal pred8x8l_horizontal_up_10, 4, 4, 6 _______________________________________________ ffmpeg-cvslog mailing list ffmpeg-cvslog@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
