Closed Bug 1827889 Opened 2 years ago Closed 2 years ago

19.74 - 5.74% reddit fcp / reddit-billgates-post-1.billg PerceptualSpeedIndex + 18 more (Linux, OSX, Windows) regression on Thu April 6 2023

Categories

(Core :: Graphics, defect)

defect

Tracking

()

RESOLVED INVALID
Tracking Status
firefox-esr102 --- unaffected
firefox112 --- unaffected
firefox113 --- affected
firefox114 --- affected

People

(Reporter: bacasandrei, Unassigned)

References

(Regression)

Details

(Keywords: perf, perf-alert, regression)

Perfherder has detected a browsertime performance regression from push b1b50f07c34b820350697dcbc6086f8b4c336be1. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

Ratio Test Platform Options Absolute values (old vs new)
20% reddit fcp macosx1015-64-shippable-qr cold fission webrender 127.19 -> 152.29
19% reddit-billgates-post-1.comments fcp windows10-64-shippable-qr cold fission webrender 170.21 -> 202.50
19% reddit-billgates-post-1.posts fcp windows10-64-shippable-qr cold fission webrender 170.21 -> 202.50
19% reddit-billgates-post-1.billg fcp windows10-64-shippable-qr cold fission webrender 170.21 -> 202.50
15% reddit-billgates-post-2.billg fcp linux1804-64-shippable-qr cold fission webrender 227.13 -> 261.67
15% reddit-billgates-post-2.hot fcp linux1804-64-shippable-qr cold fission webrender 227.13 -> 261.67
15% reddit-billgates-post-2.top fcp linux1804-64-shippable-qr cold fission webrender 227.13 -> 261.67
15% reddit-billgates-post-1.billg fcp linux1804-64-shippable-qr cold fission webrender 226.57 -> 261.00
15% reddit-billgates-post-1.comments fcp linux1804-64-shippable-qr cold fission webrender 226.57 -> 261.00
15% reddit-billgates-post-1.posts fcp linux1804-64-shippable-qr cold fission webrender 226.57 -> 261.00
... ... ... ... ...
6% reddit-billgates-post-1.billg PerceptualSpeedIndex windows10-64-shippable-qr cold fission webrender 564.21 -> 598.79
6% reddit-billgates-post-1.billg loadtime windows10-64-shippable-qr cold fission webrender 1,274.10 -> 1,348.80
6% reddit-billgates-post-1.comments loadtime windows10-64-shippable-qr cold fission webrender 1,274.10 -> 1,348.80
6% reddit-billgates-post-1.posts loadtime windows10-64-shippable-qr cold fission webrender 1,274.10 -> 1,348.80
6% reddit-billgates-post-1.billg PerceptualSpeedIndex windows10-64-shippable-qr cold fission webrender 563.59 -> 595.92

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) may be backed out in accordance with our regression policy.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a sheriff to do that for you.

For more information on performance sheriffing please see our FAQ.

Flags: needinfo?(sotaro.ikeda.g)

Set release status flags based on info from the regressing bug 1826280

Hmm, it is weird. Added code is not used with current pref webgl.out-of-process.async-present.force-sync. The change used when pref webgl.out-of-process.async-present.force-sync = false. But the pref is true by default in current m-c.

Flags: needinfo?(sotaro.ikeda.g)

The regression seemed not related to bug 1826280 because of comment 2.

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → INVALID

There weren't any re-triggers for a "before" revision for this, so I added some for the biggest regression ("20% reddit fcp macosx1015-64-shippable-qr cold fission webrender 127.19 -> 152.29"). Looking at the graph now with the retriggers it's pretty clear there was no actual regression related to bug 1826280. And maybe no regression at all, just a test that behaves weird when you compare a revision with retriggers and one without retriggers.

You need to log in before you can comment on or make changes to this bug.