[cairo] Pixman ARM Performance

[Soeren Sandmann]

Jacob Bramley <Jacob.Bramley at arm.com> writes:

> I've been having a look at the performance of Pixman on ARM and have 
> noticed something strange: It seems that between about 0.12.0 and 
> 0.15.2, the performance of pixman dropped by about 1.5 times overall. 
> Turning on the Neon optimizations in the latest source drastically 
> improves a handful of tests, but most are still very slow in comparison 
> to 0.12.0 and the overall perforamnce is pretty much the same.
> 
> Are you aware of this? Do you know what might be causing it? 

If you actually mean 0.15.2, then I am not aware of anything that
would cause slowdowns for ARM or other CPUs. In fact, between those
two releases Jeff added support for a couple of ARM v6 SIMD fast
paths.

If you mean 0.15.12, then I would not be too surprised if there are a
couple of performance regressions. Some have been fixed in 0.15.18;
there are a couple of remaining ones on the map for 0.16.0.

As Chris said, if you can post the output of cairo-perf-diff for the
versions in question, that would be helpful. If there are large
performance regressions compared to 0.12.0, we'll want to look at that
before 0.16.0.

> I can provide the actual benchmark results if they'll be useful. I 
> haven't provided them here because there are a lot of numbers!

Please do post them, or put them up somewhere and post a link.


Thanks,
Soren