I know what this is and I'll have a fix on Monday. On Sat, Dec 5, 2009 at 6:09 PM, Erik Corry <[email protected]> wrote:
> 2009/12/5 Darin Fisher <[email protected]>: > > On Fri, Dec 4, 2009 at 11:58 PM, Erik Corry <[email protected]> > wrote: > >> > >> 2009/12/5 Darin Fisher <[email protected]>: > >> > > >> > > http://build.chromium.org/buildbot/perf/vista-release-dual-core/sunspider/report.html?history=150 > >> > > >> > > http://build.chromium.org/buildbot/perf/vista-release-dual-core/v8_benchmark/report.html?history=150 > >> > Observed after the update to V8 version 2.0.3. > >> > >> Not expected. > >> > >> The sinner seems to be bitops-bitwise and. That particular benchmark > >> normally indicates a problem with access to variables on the global > >> object (the benchmark has a bug so things that look like local > >> variables are actually on the global object). Our tests on the > >> standalone engine don't show accesses to variables on the global > >> object getting slower though. This needs investigation. > >> > >> > >> > http://code.google.com/p/browserscope/source/browse/trunk/categories/sunspider/static/bitops-bitwise-and.html > > > > OK... I'll file a bug about the issue. > > http://code.google.com/p/v8/issues/detail?id=543 > > > > > > >> > >> On a different note I find those graphs excessively hard to interpret. > >> The X and Y axes have no numbers on them and for some of the > >> benchmarks the Y axis covers an absurdly small range, making every > >> little change look significant. The 'baseline' changes for no > >> apparent reason and it's not clear what it represents. You can't get > >> an overview of all the benchmarks in a suite at once and you can't see > >> which of the components you are currently displaying. > > > > Yeah, there's much room for improvement. Can you open a chromium bug > > against Area-Infrastructure for these suggestions? > > > >> > >> Is the raw data for those graphs available somehow? > > > > > http://build.chromium.org/buildbot/perf/vista-release-dual-core/sunspider/ > > > >> > >> > -Darin > >> > > >> > -- > >> > v8-dev mailing list > >> > [email protected] > >> > http://groups.google.com/group/v8-dev > >> > >> -- > >> v8-dev mailing list > >> [email protected] > >> http://groups.google.com/group/v8-dev > > > > -- > > v8-dev mailing list > > [email protected] > > http://groups.google.com/group/v8-dev > > -- > v8-dev mailing list > [email protected] > http://groups.google.com/group/v8-dev > -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
