Issue metadata
Sign in to add a comment
|
10.8% regression in blink_perf.layout at 440636:440654 |
||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Dec 29 2016
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8991947476240964608
,
Dec 29 2016
=== PERF REGRESSION === === Auto-CCing suspected CL author fayang@chromium.org === Hi fayang@chromium.org, the bisect results pointed to your CL, please take a look at the results. ===== BISECT JOB RESULTS ===== Status: completed ===== SUSPECTED CL(s) ===== Subject : Add QuicStrCat. Author : fayang Commit description: Merge internal change 142661719. Merge internal change 142781026. BUG= Review-Url: https://codereview.chromium.org/2591143003 Cr-Commit-Position: refs/heads/master@{#440641} Commit : 889bd6fc33dd8941005aa74ecceb59ceb4db2684 Date : Fri Dec 23 17:12:21 2016 ===== TESTED REVISIONS ===== Revision Mean Std Dev N Good? chromium@440635 182.917 12.1963 30 good chromium@440640 178.867 11.1581 30 good chromium@440641 166.218 9.27782 30 bad <-- chromium@440642 165.722 9.01552 30 bad chromium@440643 164.593 12.363 30 bad chromium@440645 164.258 8.49472 30 bad chromium@440654 164.71 13.8765 30 bad Bisect job ran on: mac_10_11_perf_bisect Bug ID: 677459 Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.layout Test Metric: hindi-line-layout/hindi-line-layout Relative Change: 9.95% Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_10_11_perf_bisect/builds/1149 Job details: https://chromeperf.appspot.com/buildbucket_job_status/8991947476240964608 Not what you expected? We'll investigate and get back to you! https://chromeperf.appspot.com/bad_bisect?try_job_id=5833949035626496 | 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!
,
Dec 29 2016
Does this mean our QuicStrCatImpl only hurts mac but not other platforms? I would try to improve the performance. Sorry about this.
,
Dec 29 2016
I'm surprised that a change to code in net/quic/ would affect this perf test. In order for a change to net/quic/ to affect this, it would need to be talking to a server via QUIC. QUIC is disabled in Chrome by default, but it's possible that this test enables it. Where can I learn more about what blink_perf.layout does? (I'm mostly clueless about things up at the blink layer)
,
Dec 30 2016
Could easily be a bad bisect since the graph is noisy. I'll run it again.
,
Dec 30 2016
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8991858028323443408
,
Dec 30 2016
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8991857641551049120
,
Dec 30 2016
===== BISECT JOB RESULTS ===== Status: failed === Bisection aborted === The bisect was aborted because Bisect cannot identify a culprit: Bisect failed to reproduce the regression with enough confidence. Please contact the the team (see below) if you believe this is in error. ===== TESTED REVISIONS ===== Revision Mean Std Dev N Good? chromium@440600 183.056 22.9722 30 good chromium@440700 181.757 11.4006 30 bad Bisect job ran on: mac_10_11_perf_bisect Bug ID: 677459 Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.layout Test Metric: hindi-line-layout/hindi-line-layout Relative Change: 0.71% Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_10_11_perf_bisect/builds/1150 Job details: https://chromeperf.appspot.com/buildbucket_job_status/8991858028323443408 Not what you expected? We'll investigate and get back to you! https://chromeperf.appspot.com/bad_bisect?try_job_id=5828142172733440 | 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!
,
Dec 30 2016
===== BISECT JOB RESULTS ===== Status: completed ===== SUSPECTED CL(s) ===== Subject : Add QuicStrCat. Author : fayang Commit description: Merge internal change 142661719. Merge internal change 142781026. BUG= Review-Url: https://codereview.chromium.org/2591143003 Cr-Commit-Position: refs/heads/master@{#440641} Commit : 889bd6fc33dd8941005aa74ecceb59ceb4db2684 Date : Fri Dec 23 17:12:21 2016 ===== TESTED REVISIONS ===== Revision Mean Std Dev N Good? chromium@440600 184.98 18.9673 30 good chromium@440625 184.731 11.48 30 good chromium@440638 180.115 9.5653 30 good chromium@440640 179.897 11.0197 30 good chromium@440641 165.018 9.13502 30 bad <-- chromium@440644 166.9 11.5822 30 bad chromium@440650 164.959 12.1145 30 bad chromium@440700 182.674 10.9185 30 bad Bisect job ran on: mac_10_11_perf_bisect Bug ID: 677459 Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.layout Test Metric: hindi-line-layout/hindi-line-layout Relative Change: 1.25% Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/mac_10_11_perf_bisect/builds/1151 Job details: https://chromeperf.appspot.com/buildbucket_job_status/8991857641551049120 Not what you expected? We'll investigate and get back to you! https://chromeperf.appspot.com/bad_bisect?try_job_id=5821623486119936 | 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!
,
Jan 17 2017
,
Aug 16 2017
Looks like this went unaddressed. Closing since it's 8 months old at this point. Cc-ing eae, owner of blink_perf.layout as FYI.
,
Aug 16 2017
Thanks. I'll follow up. |
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by alexclarke@chromium.org
, Dec 29 2016