<div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><div>its right, just a little improve advice in below</div><pre><br>At 2014-12-10 18:59:21,aasaipriya@multicorewareinc.com wrote:
># HG changeset patch
># User Aasaipriya Chandran <a href="mailto:aasaipriya@multicorewareinc.com>># Date 1418209141 -19800># Wed Dec 10 16:29:01 2014 +0530># Node ID 0afb5685cab0ea8a12164d64684d599d5cfe3678># Parent 29489f2fc2c7edfeb56bb0f6421d19f14d25491a>asm: chroma_hpp[4x4] for colorspace i420 in avx2 improve 217c->192c>">aasaipriya@multicorewareinc.com>
># Date 1418209141 -19800
># Wed Dec 10 16:29:01 2014 +0530
># Node ID 0afb5685cab0ea8a12164d64684d599d5cfe3678
># Parent 29489f2fc2c7edfeb56bb0f6421d19f14d25491a
>asm: chroma_hpp[4x4] for colorspace i420 in avx2 improve 217c->192c
>
</a>diff -r 29489f2fc2c7 -r 0afb5685cab0 source/common/x86/ipfilter8.asm
>--- a/source/common/x86/ipfilter8.asm Tue Dec 09 12:54:40 2014 -0600
>+++ b/source/common/x86/ipfilter8.asm Wed Dec 10 16:29:01 2014 +0530
>@@ -179,6 +179,8 @@
>
> const interp4_shuf, times 2 db 0, 1, 8, 9, 4, 5, 12, 13, 2, 3, 10, 11, 6, 7, 14, 15
>
>+interp4_hpp_shuf: times 2 db 0, 1, 2, 3, 1, 2, 3, 4, 2, 3, 4, 5, 3, 4, 5, 6
</pre><pre>we can reuse tab_Tm with vpbroadcast</pre><pre>>+
> SECTION .text
>
> cextern pb_128
>@@ -1403,6 +1405,51 @@
> jnz .loop
> RET
>
>+INIT_YMM avx2
>+cglobal interp_4tap_horiz_pp_4x4, 4,6,6
>+ mov r4d, r4m
>+
>+%ifdef PIC
>+ lea r5, [tab_ChromaCoeff]
>+ vpbroadcastd m0, [r5 + r4 * 4]
>+%else
>+ vpbroadcastd m0, [tab_ChromaCoeff + r4 * 4]
>+%endif
>+ vpbroadcastd m2, [pw_1]
>+
>+ ; register map
>+ ; m0 - interpolate coeff
>+ ; m1 - shuffle order table
>+ ; m2 - constant word 1
>+
>+ dec r0
>+
>+ ; Row 0-1
>+ vbroadcasti128 m3, [r0] ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
>+ vinserti128 m3, m3, [r0 + r1], 1
>+ pshufb m3, [interp4_hpp_shuf]
>+ pmaddubsw m3, m0
>+ pmaddwd m3, m2
>+
>+ ; Row 2-3
>+ lea r0, [r0 + r1 * 2]
>+ vbroadcasti128 m4, [r0] ; [x x x x x A 9 8 7 6 5 4 3 2 1 0]
>+ vinserti128 m4, m4, [r0 + r1], 1
>+ pshufb m4, [interp4_hpp_shuf]
>+ pmaddubsw m4, m0
>+ pmaddwd m4, m2
>+
>+ packssdw m3, m4
>+ pmulhrsw m3, [pw_512]
>+ vextracti128 xm4, m3, 1
>+ packuswb xm3, xm4
>+ lea r0, [r3 * 3]
>+ movd [r2], xm3
>+ pextrd [r2+r3], xm3, 2
>+ pextrd [r2+r3*2], xm3, 1
>+ pextrd [r2+r0], xm3, 3
>+ RET
>+
> ;--------------------------------------------------------------------------------------------------------------
> ; void interp_8tap_horiz_pp_%1x%2(pixel *src, intptr_t srcStride, pixel *dst, intptr_t dstStride, int coeffIdx)
> ;--------------------------------------------------------------------------------------------------------------
>_______________________________________________
>x265-devel mailing list
>x265-devel@videolan.org
>https://mailman.videolan.org/listinfo/x265-devel
</pre></div>