[x265] [PATCH] count_nonzero asm code, reduceded code size by combining mova and packsswb

chen chenm003 at 163.com
Fri Sep 5 19:33:11 CEST 2014


right

At 2014-09-05 13:51:22,praveen at multicorewareinc.com wrote:
># HG changeset patch
># User Praveen Tiwari
># Date 1409896229 -19800
># Node ID 0b3f68d5f1699540c71ed7b75e2b0ca965fad82f
># Parent  93db2f53fe573537bcd4eb53ca3cdb69af557eb5
>count_nonzero asm code, reduceded code size by combining mova and packsswb
>
>diff -r 93db2f53fe57 -r 0b3f68d5f169 source/common/x86/pixel-util8.asm
>--- a/source/common/x86/pixel-util8.asm	Thu Sep 04 16:42:24 2014 -0700
>+++ b/source/common/x86/pixel-util8.asm	Fri Sep 05 11:20:29 2014 +0530
>@@ -1045,7 +1045,7 @@
> ; int count_nonzero(const int16_t *quantCoeff, int numCoeff);
> ;-----------------------------------------------------------------------------
> INIT_XMM ssse3
>-cglobal count_nonzero, 2,2,4
>+cglobal count_nonzero, 2,2,3
>     pxor        m0, m0
>     shr         r1d, 4
>     movd        m1, r1d
>@@ -1053,8 +1053,7 @@
> 
> .loop:
>     mova        m2, [r0 +  0]
>-    mova        m3, [r0 + 16]
>-    packsswb    m2, m3
>+    packsswb    m2, [r0 + 16]
>     add         r0, 32
>     pcmpeqb     m2, m0
>     paddb       m1, m2
>_______________________________________________
>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/20140906/b3ac549b/attachment.html>


More information about the x265-devel mailing list