[x265] [PATCH] asm: avx2 code for add_ps[8x8] for 10bpp -- 24.9x

chen chenm003 at 163.com
Mon Mar 2 23:52:48 CET 2015


It is same speed or slower than SSE2 version, I guess this size is too small for AVX2
I have sent a demo to improve SSE2 code

At 2015-03-02 16:47:23,sumalatha at multicorewareinc.com wrote:
># HG changeset patch
># User Sumalatha Polureddy<sumalatha at multicorewareinc.com>
># Date 1425286035 -19800
># Node ID 1be088c8bc675752ebfebc4fda3bad41659269a4
># Parent  a9ad4d8202796dfb78e9d180f5fdb7cc0996ea66
>asm: avx2 code for add_ps[8x8] for 10bpp -- 24.9x
>
>add_ps[  8x8]  24.97x   275.68          6882.88
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20150303/f22cbcd2/attachment.html>


More information about the x265-devel mailing list