15.27 - 12.43% cpstartup content-process-startup / cpstartup content-process-startup (OSX) regression on Tue May 11 2021
Categories
(Core :: Graphics: Text, defect)
Tracking
()
Tracking | Status | |
---|---|---|
firefox90 | --- | affected |
People
(Reporter: alexandrui, Unassigned)
References
(Regression)
Details
(4 keywords)
Perfherder has detected a talos performance regression from push 8928abb7996b8d778fccbedb6b2d59f38b25b90b. 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% | cpstartup (docs) | content-process-startup | macosx1014-64-shippable-qr | e10s stylo webrender | 178.50 -> 205.75 |
12% | cpstartup (docs) | content-process-startup | macosx1014-64-shippable-qr | e10s stylo webrender | 177.29 -> 199.33 |
Improvements:
Ratio | Suite | Test | Platform | Options | Absolute values (old vs new) |
---|---|---|---|---|---|
3% | displaylist_mutate (docs) | macosx1015-64-shippable-qr | e10s stylo webrender | 2,058.34 -> 1,987.85 | |
3% | displaylist_mutate (docs) | macosx1015-64-shippable-qr | e10s stylo webrender-sw | 2,189.72 -> 2,115.09 |
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.
Comment 1•3 years ago
|
||
This is expected, because in bug 1708821 we effectively reverted the recent macOS startup optimization in bug 1707655 when running on pre-10.15 versions of macOS.
Looking at the perfherder graph over a 30-day period, it's pretty clear that we saw an improvement on April 30 (when bug 1708821 landed), and now we have returned to the pre-1708821 state.
Note that we were only forced to revert the change for 10.14 and earlier systems. This was due to what appears to be a OS issue (the API we're using claims to be thread-safe, but we were seeing occasional crashiness when using it off-main-thread). So while it is unfortunate we had to revert the improvement, I don't think this is actionable.
Updated•3 years ago
|
Description
•