Closed Bug 1677771 Opened 4 years ago Closed 4 years ago

5.02 - 15.22% Explicit Memory / Heap Unclassified / Resident Memory (linux1804-64-shippable-qr, macosx1014-64-shippable-qr) regression on push 6319f8438a3a47519479f0c96e6d69b420652d02 (Fri November 13 2020)

Categories

(Core :: Graphics: WebRender, defect)

Firefox 85
defect

Tracking

()

RESOLVED FIXED
85 Branch
Tracking Status
firefox-esr78 --- unaffected
firefox83 --- unaffected
firefox84 --- fixed
firefox85 --- fixed

People

(Reporter: alexandrui, Assigned: nical)

References

(Regression)

Details

(Keywords: perf, perf-alert, regression)

Attachments

(1 file)

Perfherder has detected a awsy performance regression from push 6319f8438a3a47519479f0c96e6d69b420652d02. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio Suite Test Platform Options Absolute values (old vs new)
15% Heap Unclassified linux1804-64-shippable-qr 258,750,465.88 -> 298,144,946.46
13% Resident Memory macosx1014-64-shippable-qr 715,788,069.91 -> 809,697,060.05
9% Explicit Memory linux1804-64-shippable-qr 520,974,186.75 -> 569,263,778.36
5% Resident Memory linux1804-64-shippable-qr 987,463,656.99 -> 1,037,043,978.88

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?(nical.bugzilla)
Component: Performance → Graphics: WebRender
Product: Testing → Core

Set release status flags based on info from the regressing bug 1674443

I'm bisecting the patches. They aren't supposed to change behavior much so once I know which one regressed things it shouldn't be too hard to spot the bug that ended caused the memory regression.

Assignee: nobody → nical.bugzilla
Flags: needinfo?(nical.bugzilla)

The bug is in https://hg.mozilla.org/integration/autoland/rev/6319f8438a3a

There's a loop over textures that tries to allocate a rectangle in one of them except that it doesn't stop looping after it's managed to allocate from one of them. Oops.

Pushed by nsilva@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/64036a891adc
Fix the texture cache allocating rectangles multiple times. r=jnicol
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → FIXED
Target Milestone: --- → 85 Branch
Regressions: 1624446

== Change summary for alert #27852 (as of Tue, 24 Nov 2020 14:16:42 GMT) ==

Improvements:

Ratio Suite Test Platform Options Absolute values (old vs new)
14% Heap Unclassified linux1804-64-shippable-qr 297,486,328.91 -> 254,553,702.44
12% Resident Memory macosx1014-64-shippable-qr 806,530,281.75 -> 708,353,712.96
10% Explicit Memory linux1804-64-shippable-qr 570,135,731.39 -> 512,560,465.39
10% Resident Memory macosx1014-64-shippable-qr 795,757,452.96 -> 716,913,380.54
6% Resident Memory linux1804-64-shippable-qr 1,031,910,890.14 -> 972,001,736.74
2% JS linux1804-64-shippable 79,527,719.62 -> 77,618,713.52

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=27852

Has Regression Range: --- → yes
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: