All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=693019

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmMWX6wkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2JHMvQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2MXKtgoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2Jz1pAoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmOO5rgsM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmK_MqAsM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg2OadugoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmInOlwgM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgmPaLvwoM


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

chromium-rel-win7-x64-dual
chromium-rel-win8-dual

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8987504734994025216

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : v8.browsing_desktop_ignition
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:v8:code_and_metadata_size_max/browse_media/browse_media_pinterest

Revision             Result                   N
chromium@449545      11691981 +- 2853958      21      good
chromium@449570      11813029 +- 3440030      21      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=browse.media.pinterest v8.browsing_desktop_ignition

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8987504734994025216

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5327021548240896


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8987495654817029472

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

Hi csharrison@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 : csharrison
  Commit : 59f21d4daf6f2476c0b9f6d598fff241e04fa2fe
  Date   : Fri Feb 10 05:45:12 2017
  Subject: Remove resource_loader.cc ScopedTrackers

Bisect Details
  Configuration: win_x64_perf_bisect
  Benchmark    : v8.browsing_desktop_ignition
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:code_and_metadata_size_max/browse_media/browse_media_pinterest
  Change       : 4.41% | 15382560.7619 -> 16061490.8571

Revision             Result                   N
chromium@449546      15382561 +- 4645203      21      good
chromium@449551      15376130 +- 2021623      9       good
chromium@449552      15748023 +- 2333604      14      good
chromium@449553      16375117 +- 1834919      14      bad       <--
chromium@449560      16079134 +- 2793906      21      bad
chromium@449582      16061491 +- 2244068      14      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.pinterest v8.browsing_desktop_ignition

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8987495654817029472

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5338552931450880


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Cannot be that CL, it just removes ScopedTracker instrumentation.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8987417820812255504

It's possible that that CL indirectly caused this apparent regression - if it made the resource loader slightly faster then maybe we pull in slightly more JS for this site during the time it's measured and we end up with more memory used for JS code. If so this is just a benchmark issue and we should wont-fix. I'll bisect once more to check.

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : v8.browsing_desktop_ignition
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:code_and_metadata_size_max/browse_media/browse_media_pinterest

Revision             Result                   N
chromium@449545      11535816 +- 5623289      21      good
chromium@449570      11535911 +- 5114293      21      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=browse.media.pinterest v8.browsing_desktop_ignition

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8987417820812255504

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5318239883624448


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Do you think the ScopedTrackers really cause enough overhead to change things? I thought it's essentially 2 TimeTicks::Now()s and a histogram bump.

Thanks for re-running the bisect.

Yeah it's unlikely unless the ScopedTrackers were really fine grained.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982620161657944544

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982620045769146560