Issue metadata
Sign in to add a comment
|
11.5%-33.1% regression in blink_perf.layout at 591938:591956 |
||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Sep 19
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/11def23f640000
,
Sep 20
📍 Found significant differences after each of 5 commits. https://pinpoint-dot-chromeperf.appspot.com/job/11def23f640000 Call capture event listeners in capturing phase at shadow hosts by hayato@chromium.org https://chromium.googlesource.com/chromium/src/+/da8cf24a69567fbc9d2a594489d0ec0566a3d3ad 781.4 → 777.2 (-4.24) media: Support updated CDM_11 that supports ColorSpace by xhwang@chromium.org https://chromium.googlesource.com/chromium/src/+/106f5d97d7bb58fcaf47194df98d078506be3aaf 777.2 → 785.5 (+8.36) Android: Unify android gpu crash counts by wnwen@chromium.org https://chromium.googlesource.com/chromium/src/+/d52718f19d2052f9a0c6785a091ab178b87eca66 785.9 → 769.5 (-16.43) Change NGPaintFragment::next_for_same_layout_object_ to raw pointer by yosin@chromium.org https://chromium.googlesource.com/chromium/src/+/3d757b478829a8fb67d211f932d2a91af5a51ef5 769.9 → 796.5 (+26.58) Reland: Connect Linux Font Unique Name Lookup to Blink by drott@chromium.org https://chromium.googlesource.com/chromium/src/+/f973726bdfa204cd47b1ed97070d6b54e113c231 796.4 → 782.3 (-14.11) Understanding performance regressions: http://g.co/ChromePerformanceRegressions Benchmark documentation link: https://bit.ly/blink-perf-benchmarks
,
Sep 21
My patch affects only LayoutNG. It should not cause this regression. Change NGPaintFragment::next_for_same_layout_object_ to raw pointer by yosin@chromium.org https://chromium.googlesource.com/chromium/src/+/3d757b478829a8fb67d211f932d2a91af5a51ef5
,
Sep 21
This CL shouldn't be the cause because the change is guarded by the flag. https://chromium.googlesource.com/chromium/src/+/da8cf24a69567fbc9d2a594489d0ec0566a3d3ad
,
Sep 21
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/1092e1df640000
,
Sep 21
My change https://chromium.googlesource.com/chromium/src/+/f973726bdfa204cd47b1ed97070d6b54e113c231 is only active in a unit test so far as the font unique name lookup feature is also behind a flag at this point. I don't think it can affect the blink_perf layout code.
,
Sep 22
📍 Found significant differences after each of 5 commits. https://pinpoint-dot-chromeperf.appspot.com/job/1092e1df640000 Call capture event listeners in capturing phase at shadow hosts by hayato@chromium.org https://chromium.googlesource.com/chromium/src/+/da8cf24a69567fbc9d2a594489d0ec0566a3d3ad 804.1 → 799.2 (-4.925) media: Support updated CDM_11 that supports ColorSpace by xhwang@chromium.org https://chromium.googlesource.com/chromium/src/+/106f5d97d7bb58fcaf47194df98d078506be3aaf 799.2 → 832.1 (+32.87) Android: Unify android gpu crash counts by wnwen@chromium.org https://chromium.googlesource.com/chromium/src/+/d52718f19d2052f9a0c6785a091ab178b87eca66 808.5 → 793.2 (-15.32) Change NGPaintFragment::next_for_same_layout_object_ to raw pointer by yosin@chromium.org https://chromium.googlesource.com/chromium/src/+/3d757b478829a8fb67d211f932d2a91af5a51ef5 792.2 → 843.4 (+51.2) Reland: Connect Linux Font Unique Name Lookup to Blink by drott@chromium.org https://chromium.googlesource.com/chromium/src/+/f973726bdfa204cd47b1ed97070d6b54e113c231 819.3 → 798.2 (-21.13) Understanding performance regressions: http://g.co/ChromePerformanceRegressions Benchmark documentation link: https://bit.ly/blink-perf-benchmarks
,
Jan 7
Removing myself as owner and setting to Untriaged - for blink_perf.layout pinpoint seems to find my CL beneficial, not a regression (-21.13, -14.11 respectively). Any ideas, xhwang@?
,
Jan 9
My change has no way to cause this regression since CDM_11 is no enabled yet.
,
Jan 9
GPU crash counts being accounted differently is unrelated to blink perf, removing myself.
,
Jan 11
Available, but no owner or component? Please find a component, as no one will ever find this without one. |
|||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Sep 19