Thanks for the information, Moritz! We may need to do a little bit more digging before a merge. I don't want there to be any significant performance decrease for the common single-threaded case. We'll have to figure out where the slowdown is happening and eliminate the problem.
thanks, --Andrew Whitworth On Tue, Oct 25, 2011 at 6:31 AM, Moritz Lenz <[email protected]> wrote: > Hi all, > > I've tested rakudo on top of the green_threads branch. > The good news is that all tests pass. > The bad news is that the first (and so far only) benchmark I've tried is > slower than on master. > > As a benchmark I used > https://raw.github.com/colomon/mandelbrot/master/bin/mandelbrot-color.pl > with command line argument 201 > > Analyzing script and raw results can be found at > https://gist.github.com/1312199 > > The output is: > Original: 33.965 ± 0.0858943537143158 > Patched: 36.1474 ± 0.0471445436927753 > Patched version is slower than the original by 6.43%, signficance 23.2σ > > (here "patched" refers to green_threads commit RELEASE_3_9_0-110-g807df71 > and "original" to master commit RELEASE_3_9_0-50-gec17023) > > Cheers, > Moritz > _______________________________________________ > http://lists.parrot.org/mailman/listinfo/parrot-dev > _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
