Issue metadata
Sign in to add a comment
|
11.1% regression in thread_times.tough_scrolling_cases at 484112:484174 |
||||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Jul 7 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8974716659785970416
,
Jul 9 2017
=== Auto-CCing suspected CL author chongz@chromium.org === Hi chongz@chromium.org, the bisect results pointed to your CL, please take a look at the results. === BISECT JOB RESULTS === Perf regression found with culprit Suspected Commit Author : Chong Zhang Commit : 7c75487d4702daeff63df2bd7f4a288052cd054c Date : Tue Jul 04 21:03:40 2017 Subject: Flush TouchPad generated GestureScrollEnd immediately Bisect Details Configuration: android_webview_nexus6_aosp_perf_bisect Benchmark : thread_times.tough_scrolling_cases Metric : thread_total_all_cpu_time_per_frame/thread_total_all_cpu_time_per_frame Change : 12.33% | 57.1617562773 -> 64.2101331244 Revision Result N chromium@484111 57.1618 +- 4.83924 6 good chromium@484143 57.6541 +- 2.88952 6 good chromium@484151 57.1781 +- 4.86115 6 good chromium@484155 56.8273 +- 2.06942 6 good chromium@484156 57.3986 +- 4.17457 6 good chromium@484157 64.5319 +- 0.787359 6 bad <-- chromium@484159 63.6492 +- 2.62999 6 bad chromium@484174 64.2101 +- 1.34207 6 bad To Run This Test src/tools/perf/run_benchmark -v --browser=android-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests thread_times.tough_scrolling_cases More information on addressing performance regressions: http://g.co/ChromePerformanceRegressions Debug information about this bisect: https://chromeperf.appspot.com/buildbucket_job_status/8974716659785970416 For feedback, file a bug with component Speed>Bisection
,
Jul 12 2017
,
Jul 12 2017
Should be related to issue 737379 but I'm not sure why the bisect bot points to this CL. Will see if my patch could fix this issue as well.
,
Aug 21 2017
Perf sheriff checking in. chongz: any update on this?
,
Aug 30 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969811321281483456
,
Aug 31 2017
=== Auto-CCing suspected CL author chongz@chromium.org === Hi chongz@chromium.org, the bisect results pointed to your CL, please take a look at the results. === BISECT JOB RESULTS === Perf regression found with culprit Suspected Commit Author : Chong Zhang Commit : 379f637084f5ccfe96bba687afaad8a75f53c3d5 Date : Mon Jun 26 21:23:59 2017 Subject: Re-enable VsyncAlignedInput to Field Trial Bisect Details Configuration: linux_perf_bisect Benchmark : smoothness.top_25_smooth Metric : mean_input_event_latency/Blogger Change : 11.77% | 10.3563333333 -> 11.5751666667 Revision Result N chromium@482389 10.3563 +- 0.366163 6 good chromium@482402 10.2005 +- 0.211848 6 good chromium@482404 10.3508 +- 0.301889 6 good chromium@482405 11.3428 +- 0.898892 6 bad <-- chromium@482406 11.2372 +- 0.896876 6 bad chromium@482409 11.2542 +- 0.826113 6 bad chromium@482415 11.2222 +- 1.04064 6 bad chromium@482441 11.2782 +- 0.664681 6 bad chromium@482493 11.6112 +- 0.358565 6 bad chromium@482597 11.4038 +- 0.521742 6 bad chromium@482805 11.4478 +- 0.671908 6 bad chromium@483221 11.3147 +- 0.828886 6 bad chromium@484052 11.4863 +- 0.539089 6 bad chromium@485714 11.5752 +- 0.666941 6 bad To Run This Test src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Blogger smoothness.top_25_smooth More information on addressing performance regressions: http://g.co/ChromePerformanceRegressions Debug information about this bisect: https://chromeperf.appspot.com/buildbucket_job_status/8969811321281483456 For feedback, file a bug with component Speed>Bisection
,
Aug 31 2017
,
Aug 31 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969714007565741376
,
Sep 1 2017
=== Auto-CCing suspected CL author chongz@chromium.org === Hi chongz@chromium.org, the bisect results pointed to your CL, please take a look at the results. === BISECT JOB RESULTS === Perf regression found with culprit Suspected Commit Author : Chong Zhang Commit : 379f637084f5ccfe96bba687afaad8a75f53c3d5 Date : Mon Jun 26 21:23:59 2017 Subject: Re-enable VsyncAlignedInput to Field Trial Bisect Details Configuration: linux_perf_bisect Benchmark : smoothness.top_25_smooth Metric : frame_times/Blogger Change : 4.13% | 16.6724519547 -> 17.3614761596 Revision Result N chromium@482389 16.6725 +- 0.0582644 6 good chromium@482402 16.6622 +- 0.0541652 6 good chromium@482404 16.6589 +- 0.0397812 6 good chromium@482405 17.3598 +- 0.124795 6 bad <-- chromium@482406 17.335 +- 0.115708 6 bad chromium@482409 17.3171 +- 0.28225 6 bad chromium@482415 17.3079 +- 0.273509 6 bad chromium@482441 17.3356 +- 0.170776 6 bad chromium@482493 17.3454 +- 0.17302 6 bad chromium@482597 17.3169 +- 0.0482933 6 bad chromium@482805 17.3465 +- 0.21383 6 bad chromium@483221 17.3306 +- 0.147529 6 bad chromium@484052 17.3275 +- 0.2156 6 bad chromium@485714 17.3615 +- 0.190322 6 bad To Run This Test src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Blogger smoothness.top_25_smooth More information on addressing performance regressions: http://g.co/ChromePerformanceRegressions Debug information about this bisect: https://chromeperf.appspot.com/buildbucket_job_status/8969714007565741376 For feedback, file a bug with component Speed>Bisection
,
Sep 1 2017
Merge into issue 737379 according to bisect result (bisect bot changed it back somehow). This is an expected regression since our synthetic events are generated at a rate similar to Vsync rate, thus some frames could receive no events. Note that perf bot is showing a 16ms -> 33ms regression on frame_times/Blogger, however according to trace I believe this is caused by the low rate of VSync signal (30Hz) and not related to this issue. |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Jul 7 2017