New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 880408 link

Starred by 1 user
Status: WontFix
Owner:
█ fsam...@chromium.org
Last visit > 30 days ago
Closed: Sep 5
Cc:
█ fsam...@chromium.org
maxlg@chromium.org
enne@chromium.org
jonr...@chromium.org
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

10.7%-50.5% regression in blink_perf.image_decoder at 585606:586834

Project Member Reported by maxlg@chromium.org, Sep 4 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Sep 4 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=880408

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=daf46fe1c967aa71e0b7e03a34f6456bcc918ec981e62e6e593f9218531e4f4e


Bot(s) for this bug's original alert(s):

Win 7 Nvidia GPU Perf
Win 7 Perf
mac-10_12_laptop_low_end-perf
mac-10_13_laptop_high_end-perf

blink_perf.image_decoder - Benchmark documentation link:
  https://bit.ly/blink-perf-benchmarks
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Sep 4

Cc: fsam...@chromium.org enne@chromium.org
Owner: fsam...@chromium.org
Status: Assigned (was: Untriaged)
📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/14121325640000

Revert "cc: Finish frame early if main thread is idle regardless of redraw state" by enne@chromium.org
https://chromium.googlesource.com/chromium/src/+/2e5401607fcfe8d434f2f91ae37a76222d9df322
1930 → 2076 (+146.3)

Mac OOP-D: Enable OOP-D for Mac Waterfall by fsamuel@chromium.org
https://chromium.googlesource.com/chromium/src/+/00ec720decf72b496b1b5bd5bd16bcc85d04ec03
2082 → 2412 (+330.3)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Comment 5 by fsam...@chromium.org, Sep 4 

Turning on OOP-D for Mac should not affect windows. We are running a pinpoint on Windows to better isolate where the problem is.

Comment 6 by jonr...@chromium.org, Sep 4 

Both regressions were a spike up, which has since stabilized to previous levels. With neither of the changes currently being blamed altered in the range where stabilization occurred.

Comment 7 by jonr...@chromium.org, Sep 4

Cc: jonr...@chromium.org
Project Member

Comment 8 by 42576172...@developer.gserviceaccount.com, Sep 5 

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12cec713640000

Revert "cc: Finish frame early if main thread is idle regardless of redraw state" by enne@chromium.org
https://chromium.googlesource.com/chromium/src/+/2e5401607fcfe8d434f2f91ae37a76222d9df322
1000 → 1090 (+90.37)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Comment 9 by enne@chromium.org, Sep 5

Status: WontFix (was: Assigned)
This is a revert of a test patch.  sunnyps is relanding a very similar https://chromium-review.googlesource.com/c/chromium/src/+/1205101
►

Sign in to add a comment