[x265] Fwd: [PATCH] blockcopy_sp_4x8, optimized asm code

[chen]

>diff -r 44ac213169c9 -r 1e6bf52b6e34 source/common/x86/blockcopy8.asm
>--- a/source/common/x86/blockcopy8.asm Fri Nov 08 14:46:07 2013 +0530
>+++ b/source/common/x86/blockcopy8.asm Fri Nov 08 15:04:10 2013 +0530
>@@ -948,45 +948,42 @@
> ; void blockcopy_sp_4x8(pixel *dest, intptr_t destStride, int16_t *src, intptr_t srcStride)
> ;-----------------------------------------------------------------------------
> INIT_XMM sse2
>-cglobal blockcopy_sp_4x8, 4, 6, 8, dest, destStride, src, srcStride
>+cglobal blockcopy_sp_4x8, 4, 4, 8, dest, destStride, src, srcStride

>>you have used r5
Min, r5 was in old code I have removed that. I think you are talking about [ -lea        r5,      [r4 + 2 * r3] ]. In new code I have used just 4 registers.
 
I means you have used r5, so you have to declare use 6 registers in your head
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131108/e85fa99b/attachment.html>