1505290 - 41.73% tsvgx (linux64-qr) regression on push 7a308f25794fefdd541d3c913fb807c6ee495615 (Tue Nov 6 2018)

Status: RESOLVED WORKSFORME

(Core :: Graphics: WebRender, defect, P3)

Description

[Ionuț Goldan [:igoldan]]

Talos has detected a Firefox performance regression from push:

https://hg.mozilla.org/integration/autoland/pushloghtml?changeset=7a308f25794fefdd541d3c913fb807c6ee495615

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

Regressions:

 42%  tsvgx linux64-qr opt e10s stylo     245.27 -> 347.63

Improvements:

 23%  tsvgr_opacity linux64-qr opt e10s stylo     155.84 -> 120.12
  7%  tsvg_static linux64-qr opt e10s stylo       68.62 -> 64.14


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=17389

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/Performance_sheriffing/Talos/Tests

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

Comment 1

[Ionuț Goldan [:igoldan]]

I understand this Linux QR is not a priority. Still, if you file any bugs against this issue, do link them here also.

Comment 2

[Ionuț Goldan [:igoldan]]

We also detected these new Raptor regressions:

== Change summary for alert #17398 (as of Tue, 06 Nov 2018 19:46:18 GMT) ==

Regressions:

  9%  raptor-motionmark-animometer-firefox linux64-qr opt           42.55 -> 38.87
  2%  raptor-motionmark-animometer-firefox windows10-64-qr opt      39.84 -> 38.87

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=17398

Comment 3

[Matt Woodrow (:mattwoodrow)]

A bit surprising that we regressed tsvgx by this much, especially since it improved on Windows.

I'm still looking into tsvgx regressions, so hopefully this gets better again.

Comment 4

[Maire Reavy [:mreavy]]

P4 until we understand the cause

Comment 5

[Maire Reavy [:mreavy]]

(In reply to Maire Reavy [:mreavy] Plz needinfo from comment #4)
> P4 until we understand the cause

The comment was meant for another bug

Comment 6

[Ionuț Goldan [:igoldan]]

(In reply to Matt Woodrow (:mattwoodrow) from comment #3)
> A bit surprising that we regressed tsvgx by this much, especially since it
> improved on Windows.
> 
> I'm still looking into tsvgx regressions, so hopefully this gets better
> again.

If you filed a separate bug for this investigation, please link it against this one.

Comment 7

[Matt Woodrow (:mattwoodrow)]

Apologies, added the dependent bug now.

Comment 8

[Liz Henry (:lizzard)]

Happy to take a patch in nightly or even in beta 66 for this.
I'm marking it fix-optional to remove it from weekly triage.

Comment 9

[Andrew Osmond [:aosmond] (he/him)]

Looking at the graph for the last year, I think this got fixed, as we are back around the original value we regressed from (250s). You can see a jump around April 27th, 2019 that brought us back to baseline.

https://treeherder.mozilla.org/perf.html#/graphs?highlightAlerts=1&series=mozilla-central,1648455,1,1&timerange=31536000