<div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><div>
<div></div>
<div id="divNeteaseMailCard"></div>could you try r6d as buffer and compare performance?<br><br>At 2015-02-05 18:51:21,praveen@multicorewareinc.com wrote:<br>># HG changeset patch<br>># User Praveen Tiwari<br>># Date 1423133470 -19800<br>># Node ID 7105ee079176c18464982fdc15ed0e0e05f5d015<br>># Parent 18edf28d3879454f940fced1e8f5b6cb81fe041f<br>>blockcopy_pp_4x16 asm code optimization: eliminated branch instructions<br>><br>>diff -r 18edf28d3879 -r 7105ee079176 source/common/x86/blockcopy8.asm<br>>--- a/source/common/x86/blockcopy8.asm Thu Feb 05 16:05:46 2015 +0530<br>>+++ b/source/common/x86/blockcopy8.asm Thu Feb 05 16:21:10 2015 +0530<br>>@@ -183,6 +183,40 @@<br>> RET<br>> <br>> ;-----------------------------------------------------------------------------<br>>+; void blockcopy_pp_4x16(pixel* dst, intptr_t dstStride, const pixel* src, intptr_t srcStride)<br>>+;-----------------------------------------------------------------------------<br>>+INIT_XMM sse2<br>>+cglobal blockcopy_pp_4x16, 4, 6, 4<br>>+<br>>+ lea r4, [3 * r1]<br>>+ lea r5, [3 * r3]<br>>+<br>>+ movd m0, [r2]<br>>+ movd m1, [r2 + r3]<br>>+ movd m2, [r2 + 2 * r3]<br>>+ movd m3, [r2 + r5]<br>>+<br>>+ movd [r0], m0<br>>+ movd [r0 + r1], m1<br>>+ movd [r0 + 2 * r1], m2<br>>+ movd [r0 + r4], m3<br>>+<br>>+%rep 3<br>>+ lea r2, [r2 + 4 * r3]<br>>+ movd m0, [r2]<br>>+ movd m1, [r2 + r3]<br>>+ movd m2, [r2 + 2 * r3]<br>>+ movd m3, [r2 + r5]<br>>+<br>>+ lea r0, [r0 + 4 * r1]<br>>+ movd [r0], m0<br>>+ movd [r0 + r1], m1<br>>+ movd [r0 + 2 * r1], m2<br>>+ movd [r0 + r4], m3<br>>+%endrep<br>>+ RET<br>>+<br>>+;-----------------------------------------------------------------------------<br>> ; void blockcopy_pp_%1x%2(pixel* dst, intptr_t dstStride, const pixel* src, intptr_t srcStride)<br>> ;-----------------------------------------------------------------------------<br>> %macro BLOCKCOPY_PP_W4_H8 2<br>>@@ -223,7 +257,6 @@<br>> RET<br>> %endmacro<br>> <br>>-BLOCKCOPY_PP_W4_H8 4, 16<br>> BLOCKCOPY_PP_W4_H8 4, 32<br>> <br>> ;-----------------------------------------------------------------------------<br>>_______________________________________________<br>>x265-devel mailing list<br>>x265-devel@videolan.org<br>>https://mailman.videolan.org/listinfo/x265-devel<br></div></div>