[x265] [PATCH] asm: avx code for sad_x3 for 8xN for 8bpp

[chen]

At 2015-03-05 18:50:25,sumalatha at multicorewareinc.com wrote:
># HG changeset patch
># User Sumalatha Polureddy<sumalatha at multicorewareinc.com>
># Date 1425552614 -19800
># Node ID c33d84fc67e0224c3097dc8dab24aa45efa41ed2
># Parent  b12a6a607b55ac1a347ca9c0026d9eabefe5ed80
>asm: avx code for sad_x3 for 8xN for 8bpp
>
>sad_x3[  8x4]  45.16x   277.66          12538.67
>sad_x3[  8x8]  60.13x   411.58          24747.66
>sad_x3[ 8x16]  105.51x  580.64          61261.73
>sad_x3[ 8x32]  108.69x  1005.73         109317.38
>
>diff -r b12a6a607b55 -r c33d84fc67e0 source/common/x86/asm-primitives.cpp
>--- a/source/common/x86/asm-primitives.cpp	Thu Mar 05 15:00:41 2015 +0530
>+++ b/source/common/x86/asm-primitives.cpp	Thu Mar 05 16:20:14 2015 +0530
>@@ -1145,7 +1145,7 @@
> 
> void setupAssemblyPrimitives(EncoderPrimitives &p, int cpuMask) // 8bpp
> {
>-    if (cpuMask & X265_CPU_SSE2)
>+    if (cpuMask & X265_CPU_SSE2 & 0)
>     {
>         /* We do not differentiate CPUs which support MMX and not SSE2. We only check
>          * for SSE2 and then use both MMX and SSE2 functions */
>@@ -1231,7 +1231,7 @@
> 
>         p.planecopy_sp = x265_downShift_16_sse2;
>     }
>-    if (cpuMask & X265_CPU_SSSE3)
>+    if (cpuMask & X265_CPU_SSSE3 & 0)
what's this?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20150306/fc4f97cb/attachment.html>