1479797 - 2.23 - 4.3% damp (linux64-qr, windows7-32) regression on push d01677a667bc35ccd45af746d035d3daa97571ba (Mon Jul 30 2018)

Status: RESOLVED FIXED

(DevTools :: Console, defect, P1)

Description

[Ionuț Goldan [:igoldan]]

Talos has detected a Firefox performance regression from push:

https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=6cdfa85af851cd124cd535dacd0ce8e4131aeec5&tochange=d01677a667bc35ccd45af746d035d3daa97571ba

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

Regressions:

  4%  damp linux64-qr opt e10s stylo     280.20 -> 292.26
  2%  damp windows7-32 pgo e10s stylo    236.34 -> 241.62


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=14665

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 Talos jobs in a pushlog format.

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

For information on reproducing and debugging the regression, either on try or locally, see: https://wiki.mozilla.org/Buildbot/Talos/Running

*** Please let us know your plans within 3 business days, or the offending patch(es) will be backed out! ***

Our wiki page outlines the common responses and expectations: https://wiki.mozilla.org/Buildbot/Talos/RegressionBugsHandling

Comment 1

[Ionuț Goldan [:igoldan]]

Here are the Gecko profiles for damp, on Windows 7:

before: https://perf-html.io/from-url/https%3A%2F%2Fqueue.taskcluster.net%2Fv1%2Ftask%2FKdY1tywQQ9uBRlNDZ_lnzw%2Fruns%2F0%2Fartifacts%2Fpublic%2Ftest_info%2Fprofile_damp.zip

after: https://perf-html.io/from-url/https%3A%2F%2Fqueue.taskcluster.net%2Fv1%2Ftask%2FIwwT3i0uR7OYmuwAwsX3MA%2Fruns%2F0%2Fartifacts%2Fpublic%2Ftest_info%2Fprofile_damp.zip

Comment 2

[Nicolas Chevobbe [:nchevobbe]]

This was sort of expected since we now have to import the codeMirror bundle.
But some of those numbers looks pretty bad, I'll try to lower that as much as I can.

Comment 3

[Ionuț Goldan [:igoldan]]

:nchevobbe can you provide some updates here? I noticed bug 1444132 brought many perf improvements, which basically resolved the regressions mentioned here, except the `damp linux64-qr opt e10s stylo` one; this was partially fixed.

Comment 4

[Nicolas Chevobbe [:nchevobbe]]

Sure, so Bug 1444132 improves devtools as a whole, but is not fixing the regression we introduced in Bug 1470922.
However, the test that regressed the most (webconsole autocomplete) is probably because the test itself is faulty (multiple calls to a function was done, impacting the test measurements).
A patch in Bug 1479521 makes this better, but I wonder if I could put this change in its own bug.

For the other test regressions, we have filed Bug 1482798 (slow reflow), and we still have to do some investigation on the cost of importing the codeMirror bundle in the console.

Comment 5

[Ionuț Goldan [:igoldan]]

I'm now closing this bug as fixed. Summary wise, bug 1479521 managed to complete the fix for `damp linux64-qr opt e10s stylo`.