[x265] [PATCH] asm: avx2 code for sad[32x8] for 8bpp

chen chenm003 at 163.com
Thu Mar 12 00:34:35 CET 2015


same as previous (improve in sum part)

At 2015-03-11 18:50:23,sumalatha at multicorewareinc.com wrote:
># HG changeset patch
># User Sumalatha Polureddy<sumalatha at multicorewareinc.com>
># Date 1426071015 -19800
># Node ID 8dd95d229097e31d4fc0edc7873664cc24bdeb5f
># Parent  bc9f8dbe8f655fe0b1a218bb6abb8fa88536c6dd
>asm: avx2 code for sad[32x8] for 8bpp
>
>SSE3
>sad[ 32x8]  27.37x   218.86          5989.18
>
>AVX2
>sad[ 32x8]  37.41x   158.53          5930.84
>
>diff -r bc9f8dbe8f65 -r 8dd95d229097 source/common/x86/asm-primitives.cpp
>--- a/source/common/x86/asm-primitives.cpp	Wed Mar 11 15:18:06 2015 +0530
>+++ b/source/common/x86/asm-primitives.cpp	Wed Mar 11 16:20:15 2015 +0530
>@@ -1444,6 +1444,7 @@
> 
>         p.pu[LUMA_32x32].sad = x265_pixel_sad_32x32_avx2;
>         p.pu[LUMA_64x64].sad = x265_pixel_sad_64x64_avx2;
>+        p.pu[LUMA_32x8].sad = x265_pixel_sad_32x8_avx2;
> 
>         p.pu[LUMA_8x4].sad_x3 = x265_pixel_sad_x3_8x4_avx2;
>         p.pu[LUMA_8x8].sad_x3 = x265_pixel_sad_x3_8x8_avx2;
>diff -r bc9f8dbe8f65 -r 8dd95d229097 source/common/x86/sad-a.asm
>--- a/source/common/x86/sad-a.asm	Wed Mar 11 15:18:06 2015 +0530
>+++ b/source/common/x86/sad-a.asm	Wed Mar 11 16:20:15 2015 +0530
>@@ -3925,6 +3925,34 @@
>     RET
> 
> INIT_YMM avx2
>+cglobal pixel_sad_32x8, 4,4,5
>+    xorps           m0, m0
>+%assign x 0
>+%rep 4
>+    movu           m1, [r0]               ; row 0 of pix0
>+    movu           m2, [r2]               ; row 0 of pix1
>+    movu           m3, [r0 + r1]          ; row 1 of pix0
>+    movu           m4, [r2 + r3]          ; row 1 of pix1
>+
>+    psadbw         m1, m2
>+    psadbw         m3, m4
>+    paddd          m0, m1
>+    paddd          m0, m3
>+%assign x x+1
>+  %if x < 4
>+    lea     r2,     [r2 + 2 * r3]
>+    lea     r0,     [r0 + 2 * r1]
>+  %endif
>+%endrep
>+    vextracti128   xm1, m0, 1
>+    paddd          xm0, xm1
>+    pshufd         xm1, xm0, 2
>+    paddd          xm0,xm1
>+    movd           eax, xm0
>+    RET
>+
>+
>+INIT_YMM avx2
> cglobal pixel_sad_64x64, 4,4,5
>     xorps           m0, m0
> 
>_______________________________________________
>x265-devel mailing list
>x265-devel at videolan.org
>https://mailman.videolan.org/listinfo/x265-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20150312/0c9dcfc3/attachment.html>


More information about the x265-devel mailing list