Even with bug 1396539 and bug 1394804, we still don't catch significant improvements being made to the inspector. In the following talos results I ran a first reference build including bug 1396539 and bug 1394804: https://treeherder.mozilla.org/#/jobs?repo=try&revision=5689dc70e2cc6104b87678e3d609d860cfa531af Then I pushed another one, also including this two bugs, plus a bunch of recent optimizations made to the inspector: bug 1396600, bug 1397343 and bug 1397330: https://treeherder.mozilla.org/#/jobs?repo=try&revision=280605daeddeeb64edf33a0bdcb74cd5ee2f493a This end up reporting this performance numbers: https://treeherder.mozilla.org/perf.html#/comparesubtest?originalProject=try&originalRevision=5689dc70e2cc&newProject=try&newRevision=280605daeddeeb64edf33a0bdcb74cd5ee2f493a&originalSignature=edaec66500db21d37602c99daa61ac983f21a6ac&newSignature=edaec66500db21d37602c99daa61ac983f21a6ac&filter=inspector.open&framework=1 No alert, confidence is always low. But the main issue is that it only report regressions instead of improvements! The timing of the second build are all higher. I have no clue why the results are wrong. May be we miss something similar to bug 1396539??
Or could it be related to bug 1394804 comment 30 and the numbers are not as good as before? If I understand this correctly, we only take 1 run for cold and 5 runs for warm into account.
Note that if I check performances for even more patches (all my current patch queue related to performances), I see a win: https://treeherder.mozilla.org/perf.html#/comparesubtest?originalProject=try&originalRevision=5689dc70e2cc&newProject=try&newRevision=27483b916801a063a8a7536fe321715f2f53aa94&originalSignature=edaec66500db21d37602c99daa61ac983f21a6ac&newSignature=edaec66500db21d37602c99daa61ac983f21a6ac&filter=.open&showOnlyImportant=1&framework=1 Between 5 to 17% win. It includes bug 1397341, bug 1396619 and bug 1396783 in addition to the others from comment 0.
You need to log in before you can comment on or make changes to this bug.