Closed Bug 1297646 Opened 8 years ago Closed 8 years ago

2.01 - 3.53% damp / sessionrestore_no_auto_restore / tp5o Main_RSS / tp5o Private Bytes (linux64, windows7-32, windows8-64, windowsxp) regression on push 5c56991cb9a5a2e8d3fbaa826024d9b920f3cea6 (Tue Aug 23 2016)

Categories

(Core :: JavaScript: GC, defect)

51 Branch
defect
Not set
normal

Tracking

()

RESOLVED WONTFIX
Tracking Status
firefox51 --- affected

People

(Reporter: ashiue, Unassigned)

References

Details

(Keywords: perf, regression, talos-regression)

Talos has detected a Firefox performance regression from push 5c56991cb9a5a2e8d3fbaa826024d9b920f3cea6. As author of one of the patches included in that push, we need your help to address this regression.

Summary of tests that regressed:

  sessionrestore_no_auto_restore windows8-64 opt e10s: 756.67 -> 772.08 (2.04% worse)
  tp5o Main_RSS windows7-32 pgo: 165149482.94 -> 170983259.6 (3.53% worse)
  tp5o Private Bytes windows7-32 pgo: 196276647.3 -> 202706865.67 (3.28% worse)
  damp summary windows7-32 pgo e10s: 195.6 -> 201.3 (2.91% worse)
  damp summary linux64 pgo e10s: 246.82 -> 251.78 (2.01% worse)
  damp summary linux64 opt e10s: 282.41 -> 289.57 (2.53% worse)
  damp summary windows8-64 opt e10s: 246.57 -> 254.08 (3.04% worse)
  tp5o Main_RSS windowsxp opt: 218846705.02 -> 223888760.11 (2.3% worse)
  tp5o Private Bytes windowsxp opt: 200880567.47 -> 206071368.25 (2.58% worse)
  tp5o Main_RSS windowsxp pgo: 214218242.14 -> 220360853.27 (2.87% worse)
  tp5o Private Bytes windowsxp pgo: 197216216.17 -> 203606826.11 (3.24% worse)
  sessionrestore_no_auto_restore windowsxp pgo e10s: 575.71 -> 587.92 (2.12% worse)
  damp summary windows7-32 opt e10s: 260.25 -> 266.29 (2.32% worse)


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=2661

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
After doing some retriggers, this issue might be caused by :
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=9f16afa013191b065a9ac7ab5f8081a1f4731473&tochange=5c56991cb9a5a2e8d3fbaa826024d9b920f3cea6

Hi Jon, as you are the patch author, can you take a look at this and determine what is the root cause? Thanks!
Blocks: 1291351, 1293239
Flags: needinfo?(jcoppeard)
Most of these regressions are just the tests returning to their original levels following the landing and backout of bug 1293239, specifically:

 - tp5o Main_RSS windows7-32 pgo
 - damp summary windows7-32 pgo e10s
 - damp summary linux64 opt e10s
 - damp summary windows8-64 opt e10s
 - tp5o Main_RSS windowsxp opt
 - tp5o Main_RSS windowsxp pgo
 - damp summary windows7-32 opt e10s

The following improved significantly after bug 1291292 landed and then regressed slightly after the backout of bug 1293239 so are still showing an overally improvement (due to not reserving all nursery memory up front):

 - tp5o Private Bytes windows7-32 pgo
 - tp5o Private Bytes windowsxp opt

However the following regressed and then got worse after the backout:

 - sessionrestore_no_auto_restore windowsxp pgo e10s
 - sessionrestore_no_auto_restore windows8-64 opt e10s
Flags: needinfo?(jcoppeard)
I see the main_rss and private bytes corresponding to the August 11th landing of bug 1293239.  I also see the pattern for the damp tests from the change on the 18th/19th.

given that, is there much to do here?  Should we look further into the sessionrestore_no_auto_restore times?
Component: Untriaged → JavaScript: GC
Product: Firefox → Core
:jonco, is there work to do here, or should we mark this as resolved/wontfix (just for the sessionrestore_no_auto_restore tests)?
Flags: needinfo?(jcoppeard)
I don't think there's much we can do here.
Status: NEW → RESOLVED
Closed: 8 years ago
Flags: needinfo?(jcoppeard)
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.