Performance/Status Meetings/2007-June-06: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
Line 7: Line 7:
== Agenda ==
== Agenda ==


* Choosing time/frequency of meeting: 10am-11am?
* Choosing time/frequency of meeting: 10am-11am? Every week?
* Talos status  
* Generate reliable, relevant performance data (already underway as talos). Talos status update?
* Graph server status
* Graph server status
* Areas where help is needed
* Areas where help is needed
* Reducing test variance
* Reducing test variance
* New ideas


Agenda:
* Extending meeting? Maybe 10-11?
* generate reliable, relevant performance data (already underway as talos)
* expand the scope of performance testing beyond Ts/Tp/TXUL/TDHMTL
* expand the scope of performance testing beyond Ts/Tp/TXUL/TDHMTL
* reduce noise in tests to ~1% (suggested by bz, not started)
* reduce noise in tests to ~1% (suggested by bz, not started)
* run performance tests with profiling on (rhelmer set up a machine
* run performance tests with profiling on (rhelmer set up a machine for this, but jprof has problems on trunk)
for this, but jprof has problems on trunk)
* move perf tests to chrome, so we get more reliable results, and
* move perf tests to chrome, so we get more reliable results, and
can test more than just content
can test more than just content
Line 29: Line 22:
* Better reports for sheriffs to easily spot perf regressions
* Better reports for sheriffs to easily spot perf regressions
* Tracking down specific performance issues
* Tracking down specific performance issues
* stats change to track AUS usage by osversion.
* New ideas


== Action Items ==
== Action Items ==

Revision as of 01:41, 6 June 2007

« Back to Status Meetings

Participants

Action Item Update

Agenda

  • Choosing time/frequency of meeting: 10am-11am? Every week?
  • Generate reliable, relevant performance data (already underway as talos). Talos status update?
  • Graph server status
  • Areas where help is needed
  • Reducing test variance
  • expand the scope of performance testing beyond Ts/Tp/TXUL/TDHMTL
  • reduce noise in tests to ~1% (suggested by bz, not started)
  • run performance tests with profiling on (rhelmer set up a machine for this, but jprof has problems on trunk)
  • move perf tests to chrome, so we get more reliable results, and

can test more than just content

  • improve performance reporting and analyses:
  • Graph server for easy build-to-build comparisons
  • Better reports for sheriffs to easily spot perf regressions
  • Tracking down specific performance issues
  • stats change to track AUS usage by osversion.
  • New ideas

Action Items

  • Preed: Assist Robcc with BB Master instance (use ref plat)
    • preed: handed ref image to ????. Back and forth questions on what was/wasnt needed. Some issues about setting up developer instance on local machine in addition to server image. AI:preed to meet offline robcee.
  • Alice/Robcc: File bugs/bullet list of areas others could help for perf infra
    • AI:alice: collected info on personal webpage, will post to public server soon
  • Justin & rhelmer: hardware for JProf
  • Getting higher resolution timers for tests
    • AI:Damon will meet with Boris about this. Different issues on different platforms.

Other Information

Priorities for infra:

  • Get farm up with new machines reporting data for trunk
    • AI:Robcee configuring machines with Alice, expecting to have machines reporting this afternoon.
  • Generate historical baselines
  • General profile data regularly on builds
  • Getting the perf numbers more stable
  • Developing the graph server to display time spent in each module