1950293 - 15.89 - 7.6% ts_paint / startup_about_home_paint + 8 more (Linux) regression on Sun February 23 2025

Status: RESOLVED WORKSFORME

(Core :: JavaScript Engine, defect, P3)

Description

[Treeherder Bug Filer]

Perfherder has detected a talos performance regression from push c76ecabc96ee8043ee74657a51961bdb78f757ae. 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)
16%	ts_paint	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	500.92 -> 580.50
14%	ts_paint	linux1804-64-shippable-qr	e10s fission stylo webrender	532.62 -> 609.25
11%	sessionrestore	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	809.92 -> 895.00
9%	sessionrestore_no_auto_restore	linux1804-64-shippable-qr	e10s fission stylo webrender	854.08 -> 933.33
9%	sessionrestore_no_auto_restore	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	845.00 -> 923.25
9%	sessionrestore	linux1804-64-shippable-qr	e10s fission stylo webrender	857.67 -> 936.92
8%	startup_about_home_paint	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	980.58 -> 1,063.67
8%	startup_about_home_paint_cached	linux1804-64-shippable-qr	e10s fission stylo webrender	1,026.50 -> 1,110.08
8%	startup_about_home_paint_cached	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	976.75 -> 1,054.67
8%	startup_about_home_paint	linux1804-64-shippable-qr	e10s fission stylo webrender	1,029.54 -> 1,107.75

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 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.

You can run all of these tests on try with ./mach try perf --alert 44080

The following documentation link provides more information about this command.

For more information on performance sheriffing please see our FAQ.

If you have any questions, please do not hesitate to reach out to [e-mail].

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1618391

Comment 2

[Markus Stange [:mstange]]

ts_paint shows the biggest regression, going from ~530ms to ~600ms. However, with profiling I still get ~530ms!

Here's a try push with both profiling and non-profiling "talos-other" jobs: https://treeherder.mozilla.org/jobs?repo=try&revision=76ed843f32d36004c71028ac80c50552d7819958
It shows slow ts_paint without profiling and fast ts_paint with profiling.

And for reference, here's a try push with profiling before the patch: https://treeherder.mozilla.org/jobs?repo=try&revision=e6eb4f29155654a762afd444acff018ce20a9fe0

It's going to be hard to find out what's causing the delay if it doesn't happen when profiling.

Comment 3

[Ted Campbell [:tcampbell]]

I'm firing off a few more experiments but it is looking like removing the parse time causes the benchmark to slow down. If I do both the cache I/O and do a redundant parse the performance returns to baseline.

A few data points for reference:

• selfhosted.js -> 260 kB
• selfhosted.js (libz deflate level=2) -> 39kB
• selfhosted.stencil-xdr -> 257kB
• selfhosted.stencil-xdr (lz4 level=6) -> 98kB

My startup cache seems to be around 8400kB on clean builds and the file is mmap + madvise(WILLNEED).

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1618391

Comment 5

[Ted Campbell [:tcampbell]]

This was only appearing on linux and not other platforms, and doesn't reproduce with profiling or local runs. It would still be interesting to see what timing gets thrown off by making certain parts of startup faster, but I don't believe there is real impact to worry about in 137. Will keep investigating.

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

It has been over 7 days with no activity on this performance regression.

:tcampbell, since you are the author of the regressor, bug 1618391, which triggered this performance alert, could you please provide a progress update?

If this regression is something that fixes a bug, changes the baseline of the regression metrics, or otherwise will not be fixed, please consider closing it as WONTFIX. See this documentation for more information on how to handle regressions.

For additional information/help, please needinfo the performance sheriff who filed this alert (they can be found in comment #0), or reach out in #perftest, or #perfsheriffs on Element.

For more information, please visit BugBot documentation.

Comment 7

[BugBot [:suhaib / :marco/ :calixte]]

It has been over 7 days with no activity on this performance regression.

:tcampbell, since you are the author of the regressor, bug 1618391, which triggered this performance alert, could you please provide a progress update?

If this regression is something that fixes a bug, changes the baseline of the regression metrics, or otherwise will not be fixed, please consider closing it as WONTFIX. See this documentation for more information on how to handle regressions.

For additional information/help, please needinfo the performance sheriff who filed this alert (they can be found in comment #0), or reach out in #perftest, or #perfsheriffs on Element.

For more information, please visit BugBot documentation.

Comment 8

[Ted Campbell [:tcampbell]]

I think this still might be in interesting to understand more, but setting as 'backlog-deferred' since I don't believe it is worth tracking.

Comment 9

[Mayank Bansal]

Bug 1956499 brought back a lot of these numbers to pre-regression levels. Not saying that this bug should be closed
(see the as-yet untriaged perf-alert 44625)

Comment 10

[Ted Campbell [:tcampbell]]

Based on Comment 9, I think we can close this alert to avoid confusion.