Issue metadata
Sign in to add a comment
|
7.1%-41.6% regression in thread_times.key_silk_cases at 507639:507842 |
||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Oct 12 2017
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8965932541406803456
,
Oct 12 2017
=== Auto-CCing suspected CL author robertphillips@google.com === Hi robertphillips@google.com, 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 : Robert Phillips Commit : c7e66ea989006b6f21ccf4e4c33644a55b5455ec Date : Mon Oct 09 18:07:34 2017 Subject: Disable deferred proxies a different way in Chrome Bisect Details Configuration: android_webview_nexus6_aosp_perf_bisect Benchmark : thread_times.key_silk_cases Metric : thread_raster_cpu_time_per_frame/http___jsfiddle.net_xLuvC_1_show_ Change : 41.02% | 17.6044128617 -> 24.8261731304 Revision Result N chromium@507638 17.6044 +- 0.618011 6 good chromium@507651 17.5265 +- 0.821288 6 good chromium@507657 17.6699 +- 0.473263 6 good chromium@507660 17.7321 +- 1.23825 6 good chromium@507662 17.6466 +- 0.980334 6 good chromium@507662,skia@32a6924162 17.9335 +- 0.729895 6 good chromium@507662,skia@c7e66ea989 25.0447 +- 0.650954 6 bad <-- chromium@507662,skia@37387c80cc 24.46 +- 0.625296 6 bad chromium@507662,skia@c4b015ad5f 24.6399 +- 0.501767 6 bad chromium@507662,skia@df007e1a7a 24.6812 +- 0.959501 6 bad chromium@507663 24.8262 +- 0.511873 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 --story-filter=http...jsfiddle.net.xLuvC.1.show. thread_times.key_silk_cases More information on addressing performance regressions: http://g.co/ChromePerformanceRegressions Debug information about this bisect: https://chromeperf.appspot.com/buildbucket_job_status/8965932541406803456 For feedback, file a bug with component Speed>Bisection
,
Oct 13 2017
Issue 774092 has been merged into this issue.
,
Oct 31 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a92fe336f8d823fa96c5916508303f10ce30629e commit a92fe336f8d823fa96c5916508303f10ce30629e Author: Robert Phillips <robertphillips@google.com> Date: Tue Oct 31 20:36:30 2017 Reenable Skia's deferred proxies Bug: 774090 , 769898 , 769760 Change-Id: Id6ddfa82a2c16da476b2caaa93d5d3cd99aaedc8 Reviewed-on: https://chromium-review.googlesource.com/744683 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@chromium.org> Cr-Commit-Position: refs/heads/master@{#512944} [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/skia/config/SkUserConfig.h [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/mac/virtual/gpu-rasterization/images/color-profile-background-image-cross-fade-expected.png [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/mac/virtual/gpu-rasterization/images/color-profile-background-image-cross-fade-png-expected.png [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/mac/virtual/gpu-rasterization/images/color-profile-mask-image-svg-expected.png [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/win/virtual/gpu-rasterization/images/color-profile-background-image-cross-fade-expected.png [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/win/virtual/gpu-rasterization/images/color-profile-background-image-cross-fade-png-expected.png [modify] https://crrev.com/a92fe336f8d823fa96c5916508303f10ce30629e/third_party/WebKit/LayoutTests/platform/win/virtual/gpu-rasterization/images/color-profile-mask-image-svg-expected.png
,
Nov 1 2017
,
Nov 1 2017
Skia's deferred proxies were re-enabled in the CL in #5 and it appears that the majority of the perf regressions have been resolved. I don't really know why the two metrics that didn't resolve themselves: thread_raster_cpu_time_per_frame thread_IO_cpu_time_per_frame every regressed since the causal CL was GPU-only. |
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Oct 12 2017