<div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><div><br>this patch have same problem</div>
<div>auto load r3d and check r3m, it is reduce load operator</div>
<div><br>At 2015-03-29 08:52:22,dtyx265@gmail.com wrote:<br>># HG changeset patch<br>># User David T Yuen <dtyx265@gmail.com><br>># Date 1427590314 25200<br>># Node ID ea65ec5a6c969e4cee612faa4d948e3337ed72d1<br>># Parent 36d70728acc2d9d6103af7530493176c08298ded<br>>asm:intra_pred_ang4_2 improved by ~4% 134.99 -> 129.95 with nits and tweaks<br>><br>>Corrected parameter count<br>>Reduced xmm registers to 1<br>><br>>diff -r 36d70728acc2 -r ea65ec5a6c96 source/common/x86/intrapred8.asm<br>>--- a/source/common/x86/intrapred8.asm Fri Mar 27 13:16:28 2015 -0500<br>>+++ b/source/common/x86/intrapred8.asm Sat Mar 28 17:51:54 2015 -0700<br>>@@ -1320,7 +1320,7 @@<br>> ; void intraPredAng4(pixel* dst, intptr_t dstStride, pixel* src, int dirMode, int bFilter)<br>> ;-----------------------------------------------------------------------------------------<br>> INIT_XMM sse2<br>>-cglobal intra_pred_ang4_2, 3,5,3<br>>+cglobal intra_pred_ang4_2, 4,5,1<br>> lea r4, [r2 + 2]<br>> add r2, 10<br>> cmp r3m, byte 34<br>>@@ -1328,14 +1328,12 @@<br>> <br>> movh m0, [r2]<br>> movd [r0], m0<br>>- mova m1, m0<br>>- psrldq m1, 1<br>>- movd [r0 + r1], m1<br>>- mova m2, m0<br>>- psrldq m2, 2<br>>- movd [r0 + r1 * 2], m2<br>>+ psrldq m0, 1<br>>+ movd [r0 + r1], m0<br>>+ psrldq m0, 1<br>>+ movd [r0 + r1 * 2], m0<br>> lea r1, [r1 * 3]<br>>- psrldq m0, 3<br>>+ psrldq m0, 1<br>> movd [r0 + r1], m0<br>> RET<br>> <br>>_______________________________________________<br>>x265-devel mailing list<br>>x265-devel@videolan.org<br>>https://mailman.videolan.org/listinfo/x265-devel<br></div></div>