Closed Bug 1660232 Opened 4 years ago Closed 4 years ago

2.74 - 5.31% Heap Unclassified (linux1804-64-shippable) regression on push 1f45c61b2fde6cf93c2db349c2bfa1a7f17cf44b (Mon August 17 2020)

Categories

(Core :: Layout: Text and Fonts, defect)

Firefox 81
defect

Tracking

()

RESOLVED WONTFIX
Tracking Status
firefox-esr68 --- unaffected
firefox-esr78 --- unaffected
firefox79 --- unaffected
firefox80 --- unaffected
firefox81 --- affected

People

(Reporter: aesanu, Unassigned)

References

(Regression)

Details

(Keywords: perf, perf-alert, regression)

Perfherder has detected a awsy performance regression from push 1f45c61b2fde6cf93c2db349c2bfa1a7f17cf44b. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

5% Heap Unclassified linux1804-64-shippable opt tp6 93,386,392.95 -> 98,344,268.41
5% Heap Unclassified linux1804-64-shippable opt tp6 93,468,041.94 -> 97,834,755.16
3% Heap Unclassified linux1804-64-shippable opt 77,363,549.28 -> 79,482,607.52

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the offending patch(es) will be backed out in accordance with our regression policy.

For more information on performance sheriffing please see our FAQ.

Flags: needinfo?(jfkthame)
Component: Performance → Layout: Text and Fonts
Product: Testing → Core

This is an unsurprising regression, given that bug 1659453 involves caching additional data in the font-family record to accelerate performance. The benefit can be seen in the improvements noted in https://bugzilla.mozilla.org/show_bug.cgi?id=1659453#c5. (I'm also seeing what looks like an improvement on e.g. tp5o aljazeera.net, https://treeherder.mozilla.org/perf.html#/graphs?highlightAlerts=1&series=mozilla-central,1938555,1,1&timerange=604800.)

So IMO this is simply the expected memory use of the perf improvement, and we should WONTFIX this. Lee, does that sound OK to you?

Flags: needinfo?(jfkthame) → needinfo?(lsalzman)

I think it's acceptable.

Flags: needinfo?(lsalzman)

== Change summary for alert #26767 (as of Fri, 21 Aug 2020 05:04:34 GMT) ==

Regressions:

4% raptor-tp6-bing-firefox-cold fcp linux64-shippable-qr opt webrender 277.40 -> 289.67
4% raptor-tp6-bing-firefox-cold fcp linux64-shippable-qr opt webrender 277.83 -> 289.17
3% raptor-tp6-bing-firefox-cold linux64-shippable-qr opt webrender 269.59 -> 278.04
3% raptor-tp6-yandex-firefox-cold fcp linux64-shippable-qr opt webrender 305.80 -> 314.67
2% raptor-tp6-yahoo-mail-firefox-cold fcp linux64-shippable-qr opt webrender 381.70 -> 390.92

Improvements:

41% raptor-tp6-youtube-firefox-cold fcp macosx1014-64-shippable opt 1,382.33 -> 819.83
24% raptor-tp6-linkedin-firefox-cold fcp macosx1014-64-shippable opt 1,198.95 -> 909.67
15% raptor-tp6-fandom-firefox-cold fcp linux64-shippable-qr opt webrender 713.80 -> 607.08
14% raptor-tp6-fandom-firefox-cold fcp macosx1014-64-shippable opt 1,009.45 -> 872.42
8% raptor-tp6-fandom-firefox-cold linux64-shippable-qr opt webrender 866.46 -> 795.14

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=26767

Set release status flags based on info from the regressing bug 1659453

Resolving as WONTFIX. We do see some small regressions, but also some (much more substantial) improvements. This is in line with the expected result of the patch that was landed.

Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → WONTFIX
Has Regression Range: --- → yes
You need to log in before you can comment on or make changes to this bug.