It seems like we can lower benchmark.timeToRun from 500 to 100 and still get a pretty decent number.
e.g. we spend 121s in DOM/CreateNodes now: Running DOM/CreateNodes.html (5 of 44) RESULT DOM: CreateNodes= 377.584043095 ms median= 377.545919724 ms, stdev= 1.48541786367 ms, min= 373.933014354 ms, max= 381.514733804 ms Finished: 121.340621 s By reducing the value, we only spend 30s: Running DOM/CreateNodes.html (3 of 10) RESULT DOM: CreateNodes= 376.635416667 ms median= 376.0625 ms, stdev= 2.55494737253 ms, min= 373.0 ms, max= 381.458333333 ms Finished: 29.826671 s
Created attachment 129793 [details] Reduce benchmark.timeToRun from 500ms to 100ms
Comment on attachment 129793 [details] Reduce benchmark.timeToRun from 500ms to 100ms View in context: https://bugs.webkit.org/attachment.cgi?id=129793&action=review I'm not yet a reviewer but this looks good to me. > PerformanceTests/DOM/resources/dom-perf.js:271 > + } this indentation still looks wrong.
Comment on attachment 129793 [details] Reduce benchmark.timeToRun from 500ms to 100ms View in context: https://bugs.webkit.org/attachment.cgi?id=129793&action=review >> PerformanceTests/DOM/resources/dom-perf.js:271 >> + } > > this indentation still looks wrong. Oops, yeah this isn't right :(
Comment on attachment 129793 [details] Reduce benchmark.timeToRun from 500ms to 100ms LGTM. You should consider adidng a comment next to the value about how/why 100ms was chosen. If someone had comemnted next to teh 500ms this review would have been even easier. :)
Thanks for the review. (In reply to comment #5) > (From update of attachment 129793 [details]) > LGTM. You should consider adidng a comment next to the value about how/why 100ms was chosen. If someone had comemnted next to teh 500ms this review would have been even easier. :) Done.
Landed in r109615.