[x265] [PATCH] copy_cnt: enable avx2 version of asm code

Deepthi Nandakumar deepthi at multicorewareinc.com
Thu Sep 11 08:20:14 CEST 2014


Ok, thanks.

On Thu, Sep 11, 2014 at 11:47 AM, Praveen Tiwari <
praveen at multicorewareinc.com> wrote:

> You can push 16x16 and 32x32 also they are good in performance but they
> need a bit more improvement, I will be sending improvement patch soon.
>
> Regards,
> Praveen Tiwari
>
> On Thu, Sep 11, 2014 at 11:29 AM, Deepthi Nandakumar <
> deepthi at multicorewareinc.com> wrote:
>
>> Would be better to combine this asm enable with the corresponding asm
>> patch itself. I have pushed copy_cnt8, and enabled only that for now.
>>
>> On Wed, Sep 10, 2014 at 3:28 PM, <praveen at multicorewareinc.com> wrote:
>>
>>> # HG changeset patch
>>> # User Praveen Tiwari
>>> # Date 1410343073 -19800
>>> # Node ID 2cd4a13086740728559fde3a176953e9aa4c0782
>>> # Parent  7bc4db02ccc728f6e2ddedd036c96e3d37b90f22
>>> copy_cnt: enable avx2 version of asm code
>>>
>>> diff -r 7bc4db02ccc7 -r 2cd4a1308674 source/common/x86/asm-primitives.cpp
>>> --- a/source/common/x86/asm-primitives.cpp      Wed Sep 10 14:45:33 2014
>>> +0530
>>> +++ b/source/common/x86/asm-primitives.cpp      Wed Sep 10 15:27:53 2014
>>> +0530
>>> @@ -1724,14 +1724,10 @@
>>>          p.sad_x4[LUMA_16x32] = x265_pixel_sad_x4_16x32_avx2;
>>>          p.ssd_s[BLOCK_32x32] = x265_pixel_ssd_s_32_avx2;
>>>
>>> -        /* Need to update assembly code as per changed interface of the
>>> copy_cnt primitive, once
>>> -         * code is updated, avx2 version will be enabled */
>>> -/*
>>>          p.copy_cnt[BLOCK_4x4] = x265_copy_cnt_4_avx2;
>>>          p.copy_cnt[BLOCK_8x8] = x265_copy_cnt_8_avx2;
>>>          p.copy_cnt[BLOCK_16x16] = x265_copy_cnt_16_avx2;
>>>          p.copy_cnt[BLOCK_32x32] = x265_copy_cnt_32_avx2;
>>> -*/
>>>
>>>          p.cvt32to16_shl[BLOCK_4x4] = x265_cvt32to16_shl_4_avx2;
>>>          p.cvt32to16_shl[BLOCK_8x8] = x265_cvt32to16_shl_8_avx2;
>>> _______________________________________________
>>> x265-devel mailing list
>>> x265-devel at videolan.org
>>> https://mailman.videolan.org/listinfo/x265-devel
>>>
>>
>>
>> _______________________________________________
>> x265-devel mailing list
>> x265-devel at videolan.org
>> https://mailman.videolan.org/listinfo/x265-devel
>>
>>
>
> _______________________________________________
> 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/20140911/2d164cc9/attachment-0001.html>


More information about the x265-devel mailing list