Issue metadata
Sign in to add a comment
|
3.2%-9546.9% regression in memory.top_10_mobile at 1530653166:1531359724 |
||||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Jul 13
Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8941089671870990096
,
Jul 15
=== Auto-CCing suspected CL author khushalsagar@chromium.org === Hi khushalsagar@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 : Khushal Commit : 81a71217973e595b574783c2496629311051450f Date : Mon Jul 09 22:48:23 2018 Subject: gpu: Add memory tracing for ServiceTransferCache. Bisect Details Configuration: go-webview-phone-perf-bisect Benchmark : system_health.memory_mobile Metric : memory:webview:all_processes:reported_by_chrome:gpu:effective_size_avg/load_media/load_media_imgur Change : 9546.88% | 139264.0 -> 13434628.0 Revision Result N android-chrome@df2c4f1ee9 139264 +- 0.0 6 good android-chrome@4b327d9102 139264 +- 0.0 6 good android-chrome@96555ff9c3 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573466 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573486 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573491 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573494 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573495 139264 +- 0.0 6 good android-chrome@96555ff9c3,chromium@573496 13434628 +- 0.0 6 bad <-- android-chrome@96555ff9c3,chromium@573506 13434628 +- 0.0 6 bad android-chrome@96555ff9c3,chromium@573545 13434628 +- 0.0 6 bad android-chrome@0c23a72c68 13434628 +- 0.0 6 bad android-chrome@7b94b395a3 13434628 +- 0.0 6 bad android-chrome@cd63705090 13434628 +- 0.0 6 bad android-chrome@cb4b582a48 13434628 +- 0.0 6 bad android-chrome@3a4b7c421d 13434628 +- 0.0 6 bad Please refer to the following doc on diagnosing memory regressions: https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md 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=load.media.imgur system_health.memory_mobile More information on addressing performance regressions: http://g.co/ChromePerformanceRegressions Debug information about this bisect: https://chromeperf.appspot.com/buildbucket_job_status/8941089671870990096 For feedback, file a bug with component Speed>Bisection
,
Jul 15
This is expected. The change is adding memory tracing for a component which was not being traced anymore.
,
Jul 15
,
Jul 16
,
Jul 20
,
Jul 23
,
Jul 24
This is a pretty significant regression, what actions will result from regressions/progressions we find from tracing? Is this something that needs to be traced?
,
Jul 24
Yes, OOPR introduces a new cache for transferring large payloads from the renderer to the GPU and we should have tracing for it. This is not a real regression even in terms of reporting, the responsibility for managing image uploads which used to be in the renderer and reported by cc has simply moved to the GPU process and reported under gpu now. This reporting in the GPU was missing when OOPR was initially turned on for the perf bots and this change fills that hole. |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Jul 13