All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=766594

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=f31e4001fbeca6c8fc1b040ac53913a61ab707e253b6ac9c56af70e2daec9f37


Bot(s) for this bug's original alert(s):

chromium-rel-mac12

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968030192681657792

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

Hi kbr@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 : Kenneth Russell
  Commit : cdc786f5d2a2b538f4914ba0afbaee15634206b9
  Date   : Sat Sep 16 21:07:21 2017
  Subject: Use 8x MSAA rather than 4x MSAA when available.

Bisect Details
  Configuration: mac_10_12_perf_bisect
  Benchmark    : memory.desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:gpu:effective_size_avg/TrivialWebGLPageSharedPageState
  Change       : 58.91% | 57159365.6667 -> 90829667.3333

Revision             Result                   N
chromium@502508      57159366 +- 1945848      6      good
chromium@502511      57861659 +- 3948269      6      good
chromium@502512      58563952 +- 6279856      6      good
chromium@502513      90829667 +- 2461325      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=TrivialWebGLPageSharedPageState memory.desktop

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8968030192681657792


For feedback, file a bug with component Speed>Bisection

I'm guessing a memory regression wasn't expected?

 Issue 766617  has been merged into this issue.

 Issue 766612  has been merged into this issue.

The specific sub-test failing here is:
memory:chrome:all_processes:reported_by_chrome:gpu:effective_size_avg/TrivialWebGLPageSharedPageState

which is this file:

https://chromium.googlesource.com/chromium/src/+/master/tools/perf/page_sets/trivial_sites/trivial_webgl.html

It makes a WebGL canvas which is the size of the overall window.

It's expected that an 8xMSAA (multisampled antialiasing) back buffer will consume more video memory than a 4x MSAA buffer. Further, assuming this bot is a Mac Mini with Intel GPU (I don't know how to tell this from the perf graphs; the Telemetry-based GPU tests all output this information at the top of their stdout), the buffer will be allocated out of system memory, leading to a measurable change in memory.

This memory increase is OK, especially given that Safari's already using this setting for their WebGL content. Closing this as WontFix.