Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969739408592861840

=== BISECT JOB RESULTS ===
Bisect failed for unknown reasons

Please contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : benchmark_duration/benchmark_duration


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=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests system_health.memory_desktop

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8969739408592861840


For feedback, file a bug with component Speed>Bisection

 Issue 761451  has been merged into this issue.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969631464629927792

It looks like the return code bisect was launched on "benchmark_duration" metric for system_health.memory_desktop. I don't think this was quite the correct metric to bisect on, for a couple of reasons:

1) Even when browse:media:tumblr fails, memory_desktop still returns a value for benchmark duration, which is just the sum of how long all of the stories to took to run.
2) Because calculating the entire system_health.memory_desktop benchmark's duration requires running all of the stories, it takes about 3 hours per revision that the bisect checks. This is what went wrong in the previous bisect: it hit its 24h timeout.

I think what we want to is to pick a metric (any metric that's reliably generated by system_health.memory_desktop's browse:media:tumblr story) and 
do a return code bisect on that metric. The flakiness dashboard seems to indicate that the test fails most but not all of the time, so we want to pick a repeat count that's high enough that we think we'll see the failure at least once for each revision. I kind of arbitrarily chose 5.

I'm going to kick off a new bisect with these parameters.

Disable in the CQ now: https://chromium-review.googlesource.com/c/chromium/src/+/648248

 Issue 761394  has been merged into this issue.

=== Auto-CCing suspected CL author sunnyps@chromium.org ===

Hi sunnyps@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : Sunny Sachanandani
  Commit : c7a7949c958b4867600ca716968da430fd9b952f
  Date   : Wed Aug 30 22:21:33 2017
  Subject: cc: Flush command buffers when all tiles are done.

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:locked_size_avg/browse_media/browse_media_tumblr

Revision             Exit Code      N
chromium@498540      0 +- N/A       5      good
chromium@498602      0 +- N/A       5      good
chromium@498633      0 +- N/A       5      good
chromium@498637      0 +- N/A       5      good
chromium@498639      0 +- N/A       5      good
chromium@498640      1 +- N/A       5      bad       <--
chromium@498641      1 +- N/A       5      bad
chromium@498649      1 +- N/A       5      bad
chromium@498664      1 +- N/A       5      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=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.tumblr system_health.memory_desktop

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8969631464629927792


For feedback, file a bug with component Speed>Bisection

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b766239487acfd037b77d11575098598ed80c0f8

commit b766239487acfd037b77d11575098598ed80c0f8
Author: Charlie Andrews <charliea@chromium.org>
Date: Fri Sep 01 22:05:24 2017

Disable browse:media:tumblr on Mac 10.11 Perf

TBR=nednguyen@chromium.org,altimin@chromium.org

Bug: 760966
Change-Id: If6eb9c9f865925df89ef3df74dd49dd179468b47
Reviewed-on: https://chromium-review.googlesource.com/648248
Reviewed-by: Charlie Andrews <charliea@chromium.org>
Reviewed-by: Alexander Timin <altimin@chromium.org>
Commit-Queue: Charlie Andrews <charliea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#499324}
[modify] https://crrev.com/b766239487acfd037b77d11575098598ed80c0f8/tools/perf/page_sets/system_health/expectations.py

sunnyps: Looks like https://chromium-review.googlesource.com/c/chromium/src/+/641381 causes test failures on the perfbots. Are you investigating?

I have a fix coming up. The main tracking bug is  issue 760886  but let's keep this one open too because this is specifically about the benchmark.

 Issue 762177  has been merged into this issue.

 Issue 762179  has been merged into this issue.

 Issue 762224  has been merged into this issue.

 Issue 762193  has been merged into this issue.

 Issue 762188  has been merged into this issue.

 Issue 762187  has been merged into this issue.

 Issue 762187  has been merged into this issue.

The perf dashboard detected a lot of performance regressions, the ones in #12-18 I merged into  bug 760886 

 Issue 762181  has been merged into this issue.

 Issue 762299  has been merged into this issue.

Issue 762284 has been merged into this issue.

 Issue 762176  has been merged into this issue.

 Issue 762223  has been merged into this issue.

 Issue 762186  has been merged into this issue.

sullivan@ the underlying issue is fixed. Do we need to revert the CL in #9?

Yes, please!

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/81ef09545ccc0e3e8451a4c522b90a9480e5aa8c

commit 81ef09545ccc0e3e8451a4c522b90a9480e5aa8c
Author: Sunny Sachanandani <sunnyps@chromium.org>
Date: Wed Oct 18 23:10:30 2017

Revert "Disable browse:media:tumblr on Mac 10.11 Perf"

This reverts commit b766239487acfd037b77d11575098598ed80c0f8.

Reason for revert: The underlying issue ( crbug.com/760886 ) has been fixed so reenable

Original change's description:
> Disable browse:media:tumblr on Mac 10.11 Perf
> 
> TBR=nednguyen@chromium.org,altimin@chromium.org
> 
> Bug: 760966
> Change-Id: If6eb9c9f865925df89ef3df74dd49dd179468b47
> Reviewed-on: https://chromium-review.googlesource.com/648248
> Reviewed-by: Charlie Andrews <charliea@chromium.org>
> Reviewed-by: Alexander Timin <altimin@chromium.org>
> Commit-Queue: Charlie Andrews <charliea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#499324}

TBR=nednguyen@google.com,charliea@chromium.org,altimin@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 760966
Change-Id: I031b6d73da8f31b9fd10263b67671083409dd2d1
Reviewed-on: https://chromium-review.googlesource.com/726839
Commit-Queue: Sunny Sachanandani <sunnyps@chromium.org>
Reviewed-by: Sunny Sachanandani <sunnyps@chromium.org>
Cr-Commit-Position: refs/heads/master@{#509916}
[modify] https://crrev.com/81ef09545ccc0e3e8451a4c522b90a9480e5aa8c/tools/perf/page_sets/system_health/expectations.py

Setting defect without priority to Pri-2.