1641538 - 5.08 - 61.35% perf_reftest [singletons] coalesce-1|-2.htm (linux64-shippable|-qr, macosx1014-64-shippable, windows10-64-shippable|-qr, windows7-32-shippable) regression on push 636b70578d73752e1d1a39acb9fadf641a86b779 (Wed May 27 2020)

Status: RESOLVED WONTFIX

(Core :: CSS Parsing and Computation, defect, P3)

Description

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Perfherder has detected a talos performance regression from push 636b70578d73752e1d1a39acb9fadf641a86b779. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

61% perf_reftest coalesce-2.html linux64-shippable opt e10s stylo 16.08 -> 25.94
45% perf_reftest coalesce-2.html macosx1014-64-shippable opt e10s stylo 29.84 -> 43.38
44% perf_reftest coalesce-2.html linux64-shippable-qr opt e10s stylo 18.29 -> 26.29
41% perf_reftest coalesce-2.html windows10-64-shippable-qr opt e10s stylo 19.11 -> 26.87
38% perf_reftest coalesce-2.html windows10-64-shippable opt e10s stylo 19.20 -> 26.58
32% perf_reftest coalesce-2.html windows7-32-shippable opt e10s stylo 18.73 -> 24.81
32% perf_reftest coalesce-2.html windows7-32-shippable opt e10s stylo 18.70 -> 24.77
26% perf_reftest coalesce-1.html windows10-64-shippable-qr opt e10s stylo 51.25 -> 64.54
25% perf_reftest coalesce-1.html macosx1014-64-shippable opt e10s stylo 71.34 -> 89.20
24% perf_reftest coalesce-1.html linux64-shippable-qr opt e10s stylo 47.92 -> 59.31
23% perf_reftest coalesce-1.html linux64-shippable opt e10s stylo 45.63 -> 55.96
23% perf_reftest coalesce-1.html windows10-64-shippable opt e10s stylo 50.52 -> 61.92
17% perf_reftest coalesce-1.html windows7-32-shippable opt e10s stylo 48.54 -> 56.61
8% perf_reftest_singletons coalesce-1.html windows10-64-shippable-qr opt e10s stylo 160.28 -> 173.83
8% perf_reftest_singletons coalesce-1.html windows10-64-shippable opt e10s stylo 156.69 -> 169.90
8% perf_reftest_singletons coalesce-1.html linux64-shippable-qr opt e10s stylo 152.05 -> 163.97
8% perf_reftest_singletons coalesce-1.html linux64-shippable opt e10s stylo 144.47 -> 155.41
7% perf_reftest_singletons coalesce-2.html windows10-64-shippable opt e10s stylo 125.53 -> 134.80
7% perf_reftest_singletons coalesce-2.html linux64-shippable-qr opt e10s stylo 123.39 -> 132.37
7% perf_reftest_singletons coalesce-2.html macosx1014-64-shippable opt e10s stylo 200.32 -> 214.80
7% perf_reftest_singletons coalesce-2.html linux64-shippable opt e10s stylo 116.44 -> 124.73
7% perf_reftest_singletons coalesce-1.html macosx1014-64-shippable opt e10s stylo 243.78 -> 260.72
7% perf_reftest_singletons coalesce-2.html windows10-64-shippable-qr opt e10s stylo 129.22 -> 138.01
6% perf_reftest_singletons coalesce-1.html windows7-32-shippable opt e10s stylo 148.15 -> 156.75
5% perf_reftest_singletons coalesce-2.html windows7-32-shippable opt e10s stylo 118.98 -> 125.02

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

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

Set release status flags based on info from the regressing bug 1640843

Comment 2

[Rachel Tublitz [:rachel]]

S1 or S2 bugs need an assignee - could you find someone for this bug?

Comment 3

[Emilio Cobos Álvarez (:emilio)]

It's on my queue, though I'm not too concerned about this microbenchmark regression, it's kind of expected and it's testing the only case where my patch ends up doing more work than before :)

Comment 4

[Gian-Carlo Pascutto [:gcp]]

Emilio, based on your statement it sounds like this probably doesn't have to be an S2?

Comment 5

[Emilio Cobos Álvarez (:emilio)]

Yes, this is probably an S4. I'm not actively looking into this and I think this is fairly low priority, if worth fixing at all...

Comment 6

[Emilio Cobos Álvarez (:emilio)]

I think I'm calling this a wontfix. We've recovered part of the loss here and this is the worse case of my patch which otherwise invalidates better.