1967062 - 10%-37% regression on multiple subtests of Octane, Six-speed and Jetstream2 on 15May2025 (and some improvements too)

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P3)

Description

[Mayank Bansal]

12% regression on Octane-pdfjs

8% on six-speed-spread-literal-es5
6% on six-speed-spread-literal-es6 (The number says 1.9%, but the baseline is ~100 and the new number is 106. So a 6% regression)

Range: https://hg-edge.mozilla.org/integration/autoland/pushloghtml?fromchange=d02d6a7f5f3cbb00ee3277e4dfff07980bf12cae&tochange=c2c5b8799d019fabc51a15e9ce754abfba1567b3
Suspect: bug 933149

Comment 1

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

Set release status flags based on info from the regressing bug 933149

Comment 2

[Mayank Bansal]

Potential improvements on Windows:
15% improvement on Windows onlyon Jetstream2-Base64-sp
4% on tagcloud
1.5% on jetstream2-pdfjs

Comment 3

[Mayank Bansal]

More Linux only regressions:

37% on Jetstream2-Base64-SP
10% on Jetstream2-dateformat-xparb
13% on Jetstream2-pdfjs

Comment 4

[Matthew Gaudet (he/him) [:mgaudet]]

Am I reading this right that basically regressions are only on Linux, with progressions on windows?

(Pondering if we are better served backing out patch or returning the loop under an ifdef Linux)

Comment 5

[Greg Stoll :gstoll]

Adding needinfo for Mayank for the question in comment 4.

Comment 6

[Mayank Bansal]

(In reply to Matthew Gaudet (he/him) [:mgaudet] from comment #4)

Am I reading this right that basically regressions are only on Linux, with progressions on windows?

This is what i observe too. Windows improves, linux regresses.
Caveat:

1. Octane pdfjs is run only on Linux, so we dont know if windows regressed

• However, Jetstream2 has a very similar pdfjs test, which improved slightly on Windows. I have updated comment 2 with pdfjs numbers.

2. Six-speed is run only on Linux. AFAIK, there is no equivalent test on Windows.

Comment 7

[Matthew Gaudet (he/him) [:mgaudet]]

Attachment: Bug 1967062 - Partial Backout of "Bug 933149 - Remove PodCopy loop heuristic r?mstange r=mstange" on Linux only.

We'll remove the loop heuristic on non-linux platforms, but keep it there as
that's where it continues to be effective.

Comment 8

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

Set release status flags based on info from the regressing bug 933149

Comment 9

[Donal Meehan [:dmeehan]]

Next week is the final week of beta for Fx140
:mgaudet, will this patch land and get an uplift request in time? Any concerns?

Comment 10

[Matthew Gaudet (he/him) [:mgaudet]]

Thank you for the note! I hadn't realized this hadn't landed -- Thought I had hit the button already.

Will queue for landing now. I'll make an uplift request tomorrow.

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/9e2c244bd16f
https://hg.mozilla.org/integration/autoland/rev/a5b73f2cf907
Partial Backout of "Bug 933149 - Remove PodCopy loop heuristic r=mstange" on Linux only.

Comment 12

[Mayank Bansal]

Octane-pdfjs is ~ back to pre-regression levels.
same with six-speed-es6 and six-speed-es5

Comment 13

[Sandor Molnar[:smolnar]]

https://hg.mozilla.org/mozilla-central/rev/a5b73f2cf907

Comment 14

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

The patch landed in nightly and beta is affected.
:mgaudet, is this bug important enough to require an uplift?

• If yes, please nominate the patch for beta approval.
  • See https://wiki.mozilla.org/Release_Management/Requesting_an_Uplift for documentation on how to request an uplift.
• If no, please set status-firefox140 to wontfix.

For more information, please visit BugBot documentation.

Comment 15

[Matthew Gaudet (he/him) [:mgaudet]]

Comment on attachment 9489706 [details]
Bug 1967062 - Partial Backout of "Bug 933149 - Remove PodCopy loop heuristic r?mstange r=mstange" on Linux only.

Beta/Release Uplift Approval Request

• User impact if declined/Reason for urgency: Small performance regression on Linux.
• Is this code covered by automated tests?: Yes
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Restores behaviour from Firefox prior release.
• String changes made/needed: No.
• Is Android affected?: Yes

Comment 16

[Donal Meehan [:dmeehan]]

Comment on attachment 9489706 [details]
Bug 1967062 - Partial Backout of "Bug 933149 - Remove PodCopy loop heuristic r?mstange r=mstange" on Linux only.

Approved for 140.0b6

Comment 17

[Pulsebot]

https://github.com/mozilla-firefox/firefox/commit/f5f4c2a04d6e
https://hg.mozilla.org/releases/mozilla-beta/rev/49fe9274a7cb