On 13/10/2016 11:40, Martin Storsjö wrote:
> The speedup vs C code is around 2-7x. The numbers are quite
> inconclusive though, since the checkasm test runs multiple filterings
> on top of each other, so later rounds might end up with different
> codepaths (different decisions on which filter to apply, based
> on input pixel differences). Disabling the early-exit in the asm
> doesn't give a fair comparison either though, since the C code
> only does the necessary calcuations for each row.
If it is so inconclusive might be worth reporting what happens with real
files using the full decoder.
Beside that, the improvement is already good enough IMHO.
libav-devel mailing list