Closed Bug 1482795 Opened 4 years ago Closed 4 years ago
.5 - 10 .54% Explicit Memory / Heap Unclassified (linux64, linux64-stylo-sequential, windows10-64, windows7-32) regression on push cfc33798c721117c0d16ab8d569b59f13fe2156c (Fri Aug 10 2018)
We have detected an awsy regression from push: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?changeset=cfc33798c721117c0d16ab8d569b59f13fe2156c As author of one of the patches included in that push, we need your help to address this regression. Regressions: 11% Heap Unclassified windows7-32 pgo stylo 37,721,709.65 -> 41,697,726.47 9% Heap Unclassified windows7-32 opt stylo 37,816,808.64 -> 41,354,176.60 7% Heap Unclassified windows10-64 opt stylo 63,901,252.06 -> 68,343,841.57 6% Heap Unclassified windows10-64 pgo stylo 63,965,525.33 -> 68,052,384.28 4% Explicit Memory windows7-32 pgo stylo 241,740,099.91 -> 251,157,672.80 4% Explicit Memory windows7-32 opt stylo 241,670,639.52 -> 250,691,251.86 3% Explicit Memory windows10-64 opt stylo 327,603,067.73 -> 338,031,311.47 3% Heap Unclassified linux64-stylo-sequential opt stylo-sequential86,377,378.73 -> 88,978,840.54 3% Heap Unclassified linux64 opt stylo 87,429,843.12 -> 89,614,191.24 You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=14902 On the page above you can see an alert for each affected platform as well as a link to a graph showing the history of scores for this test. There is also a link to a treeherder page showing the jobs in a pushlog format. To learn more about the regressing test(s), please see: https://wiki.mozilla.org/AWSY/Tests
Component: General → Graphics
Product: Testing → Core
:nical the patch that caused this consisted of 2 bugs: bug 1482109 and bug 1482109. Which one is more related to these regressions?
My guess would be Bug 1482109 in which I consolidated various memory pressure observers into a single implementation but I just spotted a mistake in one of the patches that basically makes the observer skip all memory pressure events. I'll put a fix up today.
Two silly mistakes: - strcmp should have been compared with 0 (can't believe I'm still making this mistake). - The original puppet widget code handled mem pressure events for non visible widgets rather than visible ones.
Assignee: nobody → nical.bugzilla
Attachment #8999637 - Flags: review?(sotaro.ikeda.g)
Attachment #8999637 - Flags: review?(sotaro.ikeda.g) → review+
(In reply to Ionuț Goldan [:igoldan], Performance Sheriffing from comment #1) > :nical the patch that caused this consisted of 2 bugs: bug 1482109 and bug > 1482109. Which one is more related to these regressions? Correcting mistake above with right info, to avoid confusion: "this consisted of 2 bugs: bug 1482109 and bug 1469496".
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/mozilla-inbound/rev/d4a387207b7a Fix the gfx memory pressure observer. r=sotaro
awsy improvements: == Change summary for alert #15008 (as of Tue, 14 Aug 2018 23:21:29 GMT) == Improvements: 9% Heap Unclassified windows7-32 opt stylo 41,665,685.51 -> 37,926,555.17 9% Heap Unclassified windows7-32 pgo stylo 41,185,151.56 -> 37,513,583.98 6% Heap Unclassified windows10-64 opt stylo 67,963,439.07 -> 63,817,786.22 6% Heap Unclassified windows10-64 pgo stylo 67,868,899.39 -> 63,788,253.59 4% Heap Unclassified osx-10-10 opt stylo 97,034,552.56 -> 93,427,381.21 4% Explicit Memory windows10-64 opt stylo 336,029,089.54 -> 323,849,849.50 3% Resident Memory windows10-64 opt stylo 457,048,615.32 -> 444,086,354.58 2% Heap Unclassified linux64-stylo-sequential opt stylo-sequential88,210,390.37 -> 86,184,509.42 For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=15008
You need to log in before you can comment on or make changes to this bug.