Issue metadata
Sign in to add a comment
|
3.1%-7.6% regression in speedometer at 433078:433157 |
||||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Nov 29 2016
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8994654106808428960
,
Nov 29 2016
=== Auto-CCing suspected CL author v8-autoroll@chromium.org === Hi v8-autoroll@chromium.org, the bisect results pointed to your CL below as possibly causing a regression. Please have a look at this info and see whether your CL be related. ===== BISECT JOB RESULTS ===== Status: completed ===== SUSPECTED CL(s) ===== Subject : Version 5.6.326 Author : v8-autoroll Commit description: Performance and stability improvements on all platforms. Cr-Commit-Position: refs/heads/5.6.326@{#1} Cr-Branched-From: 879f6599eee6e1dfcbe9a24bf688b261c03e9558-refs/heads/master@{#41014} Commit : bdd3886218dfe76e8560eb8a18401942452ae859 Date : Wed Nov 16 00:11:09 2016 ===== TESTED REVISIONS ===== Revision Mean Std Dev N Good? chromium@433133 678.803 186.77 50 good chromium@433134 679.284 188.702 50 good chromium@433134,v8@bdd3886218 720.76 133.555 50 bad <-- chromium@433135 722.547 131.283 50 bad chromium@433137 714.266 136.736 50 bad Bisect job ran on: winx64nvidia_perf_bisect Bug ID: 669505 Test Command: src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests speedometer Test Metric: EmberJS-TodoMVC/EmberJS-TodoMVC Relative Change: 5.22% Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/winx64nvidia_perf_bisect/builds/1959 Job details: https://chromeperf.appspot.com/buildbucket_job_status/8994654106808428960 Not what you expected? We'll investigate and get back to you! https://chromeperf.appspot.com/bad_bisect?try_job_id=5829032992571392 | O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq | X | for more information addressing perf regression bugs. For feedback, | / \ | file a bug with component Tests>AutoBisect. Thank you!
,
Nov 30 2016
Any clue why it didn't step into the roll?
,
Nov 30 2016
Can you restart the bisect? Otherwise please file a separate bug for bisect not stepping into V8.
,
Nov 30 2016
+simonhatch can you take a look at why it didn't step into the roll? I'll kick off an additional bisect.
,
Nov 30 2016
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8994561278661195536
,
Nov 30 2016
===== BISECT JOB RESULTS ===== Status: started ===== TESTED REVISIONS ===== Revision Mean Std Dev N Good? chromium@433121 752.817 189.312 50 good chromium@433130 761.929 183.771 50 good chromium@433134 757.603 141.35 50 good chromium@433135 809.467 117.114 50 bad chromium@433136 809.598 113.99 50 bad chromium@433138 808.605 149.749 50 bad chromium@433154 814.434 116.29 50 bad Bisect job ran on: mac_retina_perf_bisect Bug ID: 669505 Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests speedometer Test Metric: EmberJS-TodoMVC/EmberJS-TodoMVC Relative Change: 8.18% Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_retina_perf_bisect/builds/1848 Job details: https://chromeperf.appspot.com/buildbucket_job_status/8994561278661195536 Not what you expected? We'll investigate and get back to you! https://chromeperf.appspot.com/bad_bisect?try_job_id=5337262592622592 | O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq | X | for more information addressing perf regression bugs. For feedback, | / \ | file a bug with component Tests>AutoBisect. Thank you!
,
Nov 30 2016
The commit message is a bit misleading, because it looks like that V8 roll was actually a rollback: https://chromium.googlesource.com/chromium/src/+/b3028836333034e1bffde377a017d3026069ec6e Update the V8 version for Chromium branch 56. This is a roll-back to 5.6.326. This is the most stable candidate out of the latest Canaries. If this commit results in failures in Blink please contact the Blink sheriff. They may need to make changes to Blink because of the roll-back. When in doubt please try to contact the committer and reviewers of this CL before reverting it. TBR=vogelheim@chromium.org Review-Url: https://codereview.chromium.org/2508043003 Cr-Commit-Position: refs/heads/master@{#433135} I think in this case the bisect needs to be clear about what happened, logged crbug.com/670062
,
Dec 1 2016
Ah right. Then this is simple. Also the graphs have recovered: https://chromeperf.appspot.com/report?sid=abee31cc6ef51ddd123a049fc235d39b212b63290cc5d50466bc43d94845ef88&rev=433138 1. V8 made an improvement. 2. V8 was rolled back behind the improvement for the branch (flagged here as regression). 3. After the branch V8 rolled forward and the improvement got in again. This is basically wontfix, you can mark it as such. I keep it open for minor remarks: 1. Bisect actually dived in one step and reported in comment 3: https://chromium.googlesource.com/v8/v8/+/bdd3886218dfe76e8560eb8a18401942452ae859 instead of https://chromium.googlesource.com/chromium/src/+/b3028836333034e1bffde377a017d3026069ec6e The latter one, at least as additional information would have been helpful. 2. The link in comment 3 resolves to chromium/src instead of v8/v8 which must be adjusted manually.
,
Dec 1 2016
,
Dec 5 2016
,
Apr 11 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982619841479923984 |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by hablich@chromium.org
, Nov 29 2016