[Pixman] RFC: Pixman benchmark CPU time measurement

Pekka Paalanen ppaalanen at gmail.com
Wed Jun 3 00:36:52 PDT 2015

On Tue, 2 Jun 2015 15:03:01 -0700
Bill Spitzak <spitzak at gmail.com> wrote:

> I would have the first call return 0.0 and all the others return the
> difference between current time and when that first call was done. Then
> there is no worry about accuracy of floating point. I do not think any
> callers are interested in the absolute time, only in subtracting two
> results to get an elapsed time.

OpenMP threading must be taken into account, but that'd be doable.

It would indeed allow removing the elapsed() function from my proposal.

There would be a problem with clock() wrapping around too often, but
that's an orthogonal issue.

> Not sure if cpu time is what the benchmarks want. This does not include
> blocking waiting for the X server or for the GPU or for reading files.
> Elapsed real time is probably more useful.

This is Pixman, not Cairo. There are no connections to X, no GPUs, and
the benchmarks are not reading any files during the benchmark either.


More information about the Pixman mailing list