search

watery01 / libyuv

Automatically exported from code.google.com/p/libyuv
0 stars 0 forks source link

I420ToARGB - AVX2 version #216

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Here proposed an AVX2 patch for I420ToARGB. The performance gain on my Haswell 
platform is listed below:

Was(r655 on Haswell):
I420ToARGB_Any (918 ms)
I420ToARGB_Unaligned (875 ms)
I420ToARGB_Invert (847 ms)
I420ToARGB_Opt (854 ms)

Now(on same Haswell):
I420ToARGB_Any (689 ms)
I420ToARGB_Unaligned (641 ms)
I420ToARGB_Invert (626 ms)
I420ToARGB_Opt (626 ms)

Original issue reported on code.google.com by changjun...@intel.com on 11 Apr 2013 at 12:55

Attachments:

GoogleCodeExporter commented 9 years ago 
r660 checks in initial version.
We'll need variations, such as I420ToBGRA, so keeping bug open/started.

Original comment by fbarch...@google.com on 12 Apr 2013 at 8:43

GoogleCodeExporter commented 9 years ago 
As I420ToARGB is fully AVX2 optimized, closing this bug.
Future work needed for other conversions and a port to posix.

Original comment by fbarch...@chromium.org on 23 Jul 2013 at 9:42