1473598 - 2.18 - 3.28% Heap Unclassified (linux64-stylo-sequential, windows10-64, windows7-32) regression on push 17718af0d6a8 (Wed Jul 4 2018)

Status: VERIFIED FIXED

(Firefox :: Theme, defect, P1)

Description

[Ionuț Goldan [:igoldan]]

We have detected an awsy regression from push:

https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=796893f4d2f5555c8e29658bae05b98cffad0fe1&tochange=17718af0d6a867b484c30ac979fbfda1b8a722c7

As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

  3%  Heap Unclassified windows10-64 pgo stylo     40,908,039.06 -> 42,251,010.00
  3%  Heap Unclassified windows10-64 opt stylo     40,626,553.52 -> 41,928,215.24
  2%  Heap Unclassified linux64-stylo-sequential opt stylo-sequential55,474,383.54 -> 56,731,539.68
  2%  Heap Unclassified windows7-32 opt stylo      32,143,121.01 -> 32,843,735.51


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=14196

On the page above you can see an alert for each affected platform as well as a link to a graph showing the history of scores for this test. There is also a link to a treeherder page showing the jobs in a pushlog format.

To learn more about the regressing test(s), please see: https://wiki.mozilla.org/AWSY/Tests

Comment 1

[:Paolo Amadini]

For some reason using "@import" in bug 1472616 ends up using much more memory than just the new file size.

We're investigating a test failure in bug 1473034 that, once fixed, will allow us to restore the situation we had before bug 1473034. This may take a week or two. Ionuț, is that a reasonable timeline for this test, given that we are at the beginning of the cycle? We have enough time that I don't expect we'll ship this regression to users.

Comment 2

[:Paolo Amadini]

Ah, thanks Dão, I set the dependency but it was lost in the mid-air collision.

Comment 3

[Ionuț Goldan [:igoldan]]

(In reply to :Paolo Amadini from comment #1)
> For some reason using "@import" in bug 1472616 ends up using much more
> memory than just the new file size.
> 
> We're investigating a test failure in bug 1473034 that, once fixed, will
> allow us to restore the situation we had before bug 1473034. This may take a
> week or two. Ionuț, is that a reasonable timeline for this test, given that
> we are at the beginning of the cycle? We have enough time that I don't
> expect we'll ship this regression to users.

The timeline is perfectly reasonable. Good luck with the investigation!

Comment 4

[:Paolo Amadini]

Tests: https://treeherder.mozilla.org/#/jobs?repo=try&revision=4d66350c6be26fc975b238e8e38bf7867ba13473
Screenshots: https://treeherder.mozilla.org/#/jobs?repo=try&revision=08b92728d4f4e785c81402aab667780396248ef0
Baseline: https://treeherder.mozilla.org/#/jobs?repo=try&revision=43e7ad50413ab344a1752134ca1a2b72d0be834d

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/96ecf3102f8d
Remove "browser-window.css" since the workaround isn't needed anymore. r=paolo

Comment 6

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/96ecf3102f8d

Comment 7

[Ionuț Goldan [:igoldan]]

Push from comment 5 totally fixed the regressions:

== Change summary for alert #14241 (as of Sat, 07 Jul 2018 07:49:09 GMT) ==

Improvements:

 10%  Base Content JS osx-10-10 opt stylo     7,545,070.00 -> 6,768,899.33
  4%  Heap Unclassified windows10-64 pgo stylo42,241,896.71 -> 40,694,107.09
  4%  Heap Unclassified windows7-32 pgo stylo 33,188,883.89 -> 32,011,645.12
  3%  Heap Unclassified windows10-64 opt stylo41,889,073.12 -> 40,499,112.84
  3%  Heap Unclassified linux64-stylo-sequential opt stylo-sequential56,745,980.76 -> 55,177,642.90
  2%  Heap Unclassified osx-10-10 opt stylo   71,642,876.14 -> 70,015,275.50

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=14241

Comment 8

[Mike Conley (:mconley) (:⚙️)]

\o/