On 5/20/17 3:09 PM, Luca Barbato wrote:
> From: Alexandra Hájková <[email protected]>
> 

Now with the following text:

The read_vlc implementation is not force-inlined and that causes some
large performance regression on clang. Force inlining improves it
as shown in the next patch.

x86
gcc-6.3     -1.1%
clang-4.0   -6.1%

x86_64
gcc-6.3      2.0%
clang-4.0   -0.8%

arm32
gcc-6.3     -1.5%
clang-4.0   -2.9%

arm64
gcc-6.3     1.3%
clang-4.0   -4.2%

_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to