1503503 - 0.34 - 73.3% build times / installer size (android-4-0-armv7-api16, android-5-0-aarch64, linux32, linux64) regression on push 434f70360933449c52ece7a00a6268391f805f08 (Thu Oct 25 2018)

Status: RESOLVED INVALID

(Firefox Build System :: Toolchains, defect)

Description

[Ionuț Goldan [:igoldan]]

We have detected a build metrics regression from push:

https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?changeset=434f70360933449c52ece7a00a6268391f805f08

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

Regressions:

 73%  build times android-4-0-armv7-api16 opt taskcluster-c5.4xlarge     582.44 -> 1,009.39
 68%  build times android-5-0-aarch64 opt taskcluster-c5.4xlarge         592.66 -> 992.72
 48%  build times android-5-0-aarch64 opt taskcluster-m5.4xlarge         679.25 -> 1,003.90
  0%  installer size linux32 pgo                                         66,475,818.96 -> 66,728,152.83
  0%  installer size linux64 pgo                                         66,730,444.79 -> 66,954,363.08


You can find links to graphs and comparison views for each of the above tests at: https://treeherder.mozilla.org/perf.html#/alerts?id=17136

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://developer.mozilla.org/en-US/docs/Mozilla/Performance/Automated_Performance_Testing_and_Sheriffing/Build_Metrics

Comment 1

[Ionuț Goldan [:igoldan]]

(In reply to Ionuț Goldan [:igoldan], Performance Sheriffing from comment #0)
> Regressions:
> 
>  73%  build times android-4-0-armv7-api16 opt taskcluster-c5.4xlarge    
> 582.44 -> 1,009.39
>  68%  build times android-5-0-aarch64 opt taskcluster-c5.4xlarge        
> 592.66 -> 992.72
>  48%  build times android-5-0-aarch64 opt taskcluster-m5.4xlarge        
> 679.25 -> 1,003.90

The build times aren't really that big. It's the test noise that accentuates their magnitude.

Comment 2

[Mike Hommey [:glandium]]

It seems like the clang version change made us hit more cache misses, as one would expect, and it seems that after a while the build times are back down to their old level.