At 2015-03-06 16:19:54,[email protected] wrote: ># HG changeset patch ># User Aasaipriya Chandran <[email protected]> ># Date 1425629985 -19800 ># Fri Mar 06 13:49:45 2015 +0530 ># Node ID 52f29a8cdecf7a3b5930b640c2b507c4f11b017a ># Parent 29d4d0251f06777190867d9de35c43379a414342 >chroma_hps[16x16]for i420 avx2 asm code: 1624c->1093c > >diff -r 29d4d0251f06 -r 52f29a8cdecf source/common/x86/asm-primitives.cpp >--- a/source/common/x86/asm-primitives.cpp Fri Mar 06 13:46:52 2015 +0530 >+++ b/source/common/x86/asm-primitives.cpp Fri Mar 06 13:49:45 2015 +0530 >@@ -1792,6 +1792,7 @@ > p.chroma[X265_CSP_I420].pu[CHROMA_420_16x16].filter_hpp = > x265_interp_4tap_horiz_pp_16x16_avx2; > > p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].filter_hps = > x265_interp_4tap_horiz_ps_4x4_avx2; >+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x16].filter_hps = >x265_interp_4tap_horiz_ps_16x16_avx2; > > p.pu[LUMA_4x4].luma_vps = x265_interp_8tap_vert_ps_4x4_avx2; > >diff -r 29d4d0251f06 -r 52f29a8cdecf source/common/x86/ipfilter8.asm >--- a/source/common/x86/ipfilter8.asm Fri Mar 06 13:46:52 2015 +0530 >+++ b/source/common/x86/ipfilter8.asm Fri Mar 06 13:49:45 2015 +0530 >@@ -1806,6 +1806,61 @@ > RET > > >+;----------------------------------------------------------------------------------------------------------------------------- >+; void interp_4tap_horiz_ps_16x16(pixel *src, intptr_t srcStride, int16_t >*dst, intptr_t dstStride, int coeffIdx, int isRowExt) >+;-----------------------------------------------------------------------------------------------------------------------------; >+INIT_YMM avx2 >+cglobal interp_4tap_horiz_ps_16x16, 4,7,5 >+ mov r4d, r4m >+ mov r5d, r5m >+ add r3d, r3d >+ >+%ifdef PIC >+ lea r6, [tab_ChromaCoeff] >+ vpbroadcastd m0, [r6 + r4 * 4] >+%else >+ vpbroadcastd m0, [tab_ChromaCoeff + r4 * 4] >+%endif >+ >+ vpbroadcastd m2, [pw_1]
vbroadcasti128? >+ movu m1, [tab_Tm] unaligned? >+ >+ ; register map >+ ; m0 - interpolate coeff >+ ; m1 - shuffle order table >+ ; m2 - constant word 1 >+ mov r6d, 16 >+ dec r0 >+ cmp r5m, byte 0 >+ je .label >+ sub r0 , r1 >+ add r6d , 3 >+ >+.label .loop? >+ ; Row 0 >+ vbroadcasti128 m3, [r0] ; [x x x x x >A 9 8 7 6 5 4 3 2 1 0] >+ pshufb m3, m1 >+ pmaddubsw m3, m0 >+ pmaddwd m3, m2 >+ >+ ; Row 1 >+ vbroadcasti128 m4, [r0 + 8] ; [x x x x >x A 9 8 7 6 5 4 3 2 1 0] >+ pshufb m4, m1 >+ pmaddubsw m4, m0 >+ pmaddwd m4, m2 >+ >+ packssdw m3, m4 >+ psubw m3, [pw_2000] we have more free register, so buffer the constant since here is loop >+ >+ vpermq m3, m3, 11011000b >+ movu [r2], m3 >+ >+ lea r2, [r2 + r3] >+ lea r0, [r0 + r1] add? >+ dec r6d >+ jnz .label >+ RET >+ > ;----------------------------------------------------------------------------------------------------------------------------- > ;void interp_horiz_ps_c(const pixel* src, intptr_t srcStride, int16_t* dst, > intptr_t dstStride, int coeffIdx, int isRowExt) > ;----------------------------------------------------------------------------------------------------------------------------- >_______________________________________________ >x265-devel mailing list >[email protected] >https://mailman.videolan.org/listinfo/x265-devel
_______________________________________________ x265-devel mailing list [email protected] https://mailman.videolan.org/listinfo/x265-devel
