[x265] [PATCH] assembly code for pixel_sad_x3_4x16 and pixel_sad_x4_4x16
Steve Borho
steve at borho.org
Wed Oct 30 07:57:56 CET 2013
On Wed, Oct 30, 2013 at 1:08 AM, <yuvaraj at multicorewareinc.com> wrote:
> # HG changeset patch
> # User Yuvaraj Venkatesh <yuvaraj at multicorewareinc.com>
> # Date 1383113029 -19800
> # Wed Oct 30 11:33:49 2013 +0530
> # Node ID c218021583e62a36c06d8f8787e5e5bec13beae6
> # Parent c946d617fd9fbd2ae237d5ba7d8338a1a3f5ea7d
> assembly code for pixel_sad_x3_4x16 and pixel_sad_x4_4x16
>
Well, that was easy. Thanks
In the future, any patches like this one which make intrinsic functions no
longer necessary should remove those intrinsics in the same patch.
> diff -r c946d617fd9f -r c218021583e6 source/common/x86/asm-primitives.cpp
> --- a/source/common/x86/asm-primitives.cpp Tue Oct 29 15:16:28 2013
> +0530
> +++ b/source/common/x86/asm-primitives.cpp Wed Oct 30 11:33:49 2013
> +0530
> @@ -221,8 +221,8 @@
> {
> INIT8_NAME(sse_pp, ssd, _mmx);
> INIT8(sad, _mmx2);
> - INIT7(sad_x3, _mmx2);
> - INIT7(sad_x4, _mmx2);
> + INIT8(sad_x3, _mmx2);
> + INIT8(sad_x4, _mmx2);
> INIT8(satd, _mmx2);
> HEVC_SATD(mmx2);
> p.satd[LUMA_12x16] = cmp<12, 16, 4, 16,
> x265_pixel_satd_4x16_mmx2>;
> diff -r c946d617fd9f -r c218021583e6 source/common/x86/sad-a.asm
> --- a/source/common/x86/sad-a.asm Tue Oct 29 15:16:28 2013 +0530
> +++ b/source/common/x86/sad-a.asm Wed Oct 30 11:33:49 2013 +0530
> @@ -1111,6 +1111,7 @@
> SAD_X 3, 8, 16
> SAD_X 3, 8, 8
> SAD_X 3, 8, 4
> +SAD_X 3, 4, 16
> SAD_X 3, 4, 8
> SAD_X 3, 4, 4
> SAD_X 4, 16, 16
> @@ -1118,6 +1119,7 @@
> SAD_X 4, 8, 16
> SAD_X 4, 8, 8
> SAD_X 4, 8, 4
> +SAD_X 4, 4, 16
> SAD_X 4, 4, 8
> SAD_X 4, 4, 4
>
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
--
Steve Borho
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131030/2143aabb/attachment.html>
More information about the x265-devel
mailing list