Performance/Snappy/2012-06-21
From MozillaWiki
< Performance | Snappy
Contents
- 1 Snappy June 21 Status (no meeting this week)
- 1.1 Actions
- 1.2 Incoming
- 1.3 Projects
- 1.3.1 Better DOM event/task scheduling - nathan (telemetry)
- 1.3.2 Front-end - Gavin, Jared and Frank (Australis tab strip)
- 1.3.3 Fix cache - hurley
- 1.3.4 Profiler - jrmuizel/BenWa/Ehsan (and more)
- 1.3.5 Snappy Scrolling (meta bug 710372) - Jared
- 1.3.6 Peptest - mcote, ahal
- 1.3.7 GC pause reduction - billm
- 1.3.8 CC pause reduction - smaug, mccr8 (meta bug 698919)
Snappy June 21 Status (no meeting this week)
Actions
- gavin to find an owner for persona slowness
Incoming
- Firefox does very poorly on the IE chalkboard test https://ie.microsoft.com/testdrive/Performance/Chalkboard/
- Another perf test from MS, bubbles http://ie.microsoft.com/testdrive/performance/bubbles/
Projects
Better DOM event/task scheduling - nathan (telemetry)
- Results from the past week
- Todo this week
- More debugging, implement true per-window event queues.
Front-end - Gavin, Jared and Frank (Australis tab strip)
- Results from the past week
- Jared passed on the work for Australis (bug 738491) to Frank.
- Todo this week
- Frank is in the process of changing from using images for the tab shape to linear-gradients for theming/platform-integration purposes. This elevates the necessity of bug 761393
Fix cache - hurley
- Results from the past week
- https://bugzilla.mozilla.org/show_bug.cgi?id=765665
- https://bugzilla.mozilla.org/show_bug.cgi?id=766973
- https://bugzilla.mozilla.org/show_bug.cgi?id=763342
Profiler - jrmuizel/BenWa/Ehsan (and more)
- Results from the past week
- Frame pointers on in Nightly builds. https://groups.google.com/forum/#!topic/mozilla.dev.platform/UENmwUOFCkU/discussion
- bug 763124: Lowered hang reporter threshold to 5 secs
Snappy Scrolling (meta bug 710372) - Jared
- Results from the past week
- Jared emailed Avi to see if he is still doing work on it. No other updates.
- Todo this week
- No planned work on it this week.
Peptest - mcote, ahal
- Results from the past week
- taras, mcote, and ahal revisited peptest's rationale and current implementation.
- Good points: peptest is not only measuring something that no other test suite measures (responsiveness), but it also simulates user input better than any current framework. We can also easily extend it to record the duration of particular actions as well as their responsiveness.
- Not so good points: our current tests are not that useful. They are real-world actions, but most aren't actions for which we really care about responsiveness.
- Filed bug 767048 to implement a simple, useful test case, in which we load a comlex page in 10 tabs and switch between them, measuring responsivness and how long it takes from click to end of painting.
- Also filed bug 767055 to implement a way to measure responsiveness asynchronously; that is, start measurements, do stuff possibly including waiting on callbacks, and then stop measurements.
- Todo this week
- Start work on bug 767048.
GC pause reduction - billm
- Results from the past week
- no updates
CC pause reduction - smaug, mccr8 (meta bug 698919)
- Results from the past week
- no updates