Issue metadata
Sign in to add a comment
|
Some metrics in thread_times benchmarks drop to 0 |
||||||||||||||||||||||
Issue descriptionhttps://chromeperf.appspot.com/report?sid=108f2f611fcde261c3dcdb219ac0620ac42126310b6d1bd46001b6589eee0a5e I noticed this when triaging perf bugs. Was this expected? Kicking off some bisects. Setting to Pri-1, but please drop priority if it was not expected.
,
Mar 19 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/128cbc39440000
,
Mar 19 2018
📍 Pinpoint job started. https://pinpoint-dot-chromeperf.appspot.com/job/12a66e19440000
,
Mar 19 2018
I noticed that reference builds do NOT drop to 0, so this is likely a change outside of telemetry: https://chromeperf.appspot.com/report?sid=44d052ed188d92e0ce6813ffd5cb403a71ce373e35ca4b5942da9407a691e14c
,
Mar 19 2018
📍 Found a significant difference after 1 commit. https://pinpoint-dot-chromeperf.appspot.com/job/15c6a0b9440000 [scheduler] Use WorkerSchedulerProxy in DedicatedWorkerThread. by altimin@chromium.org https://chromium.googlesource.com/chromium/src/+/053b619cf7ac1dd0ed6504e6e65da0e45d0a9d4e Understanding performance regressions: http://g.co/ChromePerformanceRegressions
,
Mar 19 2018
altimin: this CL caused a lot of the thread_times metrics to drop to 0. Was this intentional?
,
Mar 19 2018
📍 Found a significant difference after 1 commit. https://pinpoint-dot-chromeperf.appspot.com/job/12a66e19440000 [scheduler] Use WorkerSchedulerProxy in DedicatedWorkerThread. by altimin@chromium.org https://chromium.googlesource.com/chromium/src/+/053b619cf7ac1dd0ed6504e6e65da0e45d0a9d4e Understanding performance regressions: http://g.co/ChromePerformanceRegressions
,
Mar 19 2018
📍 Found a significant difference after 1 commit. https://pinpoint-dot-chromeperf.appspot.com/job/128cbc39440000 [scheduler] Use WorkerSchedulerProxy in DedicatedWorkerThread. by altimin@chromium.org https://chromium.googlesource.com/chromium/src/+/053b619cf7ac1dd0ed6504e6e65da0e45d0a9d4e Understanding performance regressions: http://g.co/ChromePerformanceRegressions
,
Mar 19 2018
Re #6: No, this is not intentional -- the compositor thread name was renamed accidentally. The fix has already landed. |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Mar 19 2018