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

Issue 816512 link

Starred by 1 user

5.9% regression in blink_perf.image_decoder at 538208:538285

Project Member Reported by tdres...@chromium.org, Feb 26 2018 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Feb 26 2018 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=816512

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=519a4f9709fd558599c7cc3ea087137b3f01a8574df623516ff6f142791ffb6e


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

android-nexus5X

Comment 4 by sullivan@chromium.org, Mar 16 2018 

Trying another bisect.
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Mar 17 2018

Cc: h...@chromium.org thakis@chromium.org
Owner: thakis@chromium.org
Status: Assigned (was: Untriaged)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/119ff4be440000

Roll clang 324578:325667. by thakis@chromium.org
https://chromium.googlesource.com/chromium/src/+/f316d5f1b672a7bb0d726610392633ae0565b129

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

Comment 7 by 42576172...@developer.gserviceaccount.com, Mar 17 2018

Cc: a...@chromium.org egdaniel@google.com robertph...@google.com backer@chromium.org jvanverth@google.com brianosman@google.com f...@opera.com lassey@chromium.org rch@chromium.org flackr@chromium.org emir...@chromium.org mkwst@chromium.org vmi...@chromium.org bsalo...@google.com herb@google.com smcgruer@chromium.org mmenke@chromium.org wangxianzhu@chromium.org chfremer@chromium.org
Owner: smcgruer@chromium.org
📍 Found significant differences after each of 9 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/13e8d4be440000

[SPv175+] SetNeedsPaintPropertyUpdate on overflow property change for SVGForeignObject and SVGViewportContainer by wangxianzhu@chromium.org
https://chromium.googlesource.com/chromium/src/+/1d3c50f6f1b3289411d1050487efd874ad84f0ca

Refactor --enable-gpu-client-logging for reuse. by backer@chromium.org
https://chromium.googlesource.com/chromium/src/+/1a55d47f4945223f686bece5201af4c3c7dcc94b

Implement DNS over HTTPS (DoH) by lassey@chromium.org
https://chromium.googlesource.com/chromium/src/+/786929ad1cfbc97962ff5672e2469460ff535f41

Roll clang 324578:325667. by thakis@chromium.org
https://chromium.googlesource.com/chromium/src/+/f316d5f1b672a7bb0d726610392633ae0565b129

Add feature to enable multiplex codec by emircan@chromium.org
https://chromium.googlesource.com/chromium/src/+/501897f108fb1a191103f309be78e7b853a572fb

Add isFinite() bounds check for drawPoints by jvanverth@google.com
https://skia.googlesource.com/skia/+/5d32b83102a4477632f0bd360b3584a144329186

Update createWrappedRenderTargetProxy with BackendTex to use lazy proxies by egdaniel@google.com
https://skia.googlesource.com/skia/+/f87651e7b8c54569ace96ce9bf1671d5401eb91c

Don't use LOD bias on OES external textures by brianosman@google.com
https://skia.googlesource.com/skia/+/c35d7eab7b35d1acf65549720188d9ac77dde05a

Partially implement KeyframeEffectReadOnly::target() by smcgruer@chromium.org
https://chromium.googlesource.com/chromium/src/+/83b0a6ecb64bfd7d0def8a1e7dc4e58e90cad4f0

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

Comment 8 by smcgruer@chromium.org, Mar 19 2018

Owner: thakis@chromium.org
My CL simply exposed an existing member function to Javascript in the Web Animations stack. I am confident it was not a factor in a performance regression.

Looking at https://pinpoint-dot-chromeperf.appspot.com/job/13e8d4be440000, my guess would be the clang roll was the most likely suspect, assigning to Thakis.
►

Sign in to add a comment