Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975518406522741520

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975518395232978992

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975518384121138704

Fixing title to match graphs, be consistent with other bugs.

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_media


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_x64 --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/8975518384121138704


For feedback, file a bug with component Speed>Bisection

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_search


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_x64 --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/8975518406522741520


For feedback, file a bug with component Speed>Bisection

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_news


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_x64 --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/8975518395232978992


For feedback, file a bug with component Speed>Bisection

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975426257653072576

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975426256563870768

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975426254740982208

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_media


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_x64 --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/8975426257653072576


For feedback, file a bug with component Speed>Bisection

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_search


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_x64 --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/8975426256563870768


For feedback, file a bug with component Speed>Bisection

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

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


Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_news


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_x64 --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/8975426254740982208


For feedback, file a bug with component Speed>Bisection

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975334844057153856

These are timing out because you're not bisecting with a story filter. You need to click the individual page in the legend (for example, browse_news_flipboard) so the dashboard knows how to set the story filter. Kicked another bisect.

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

Hi hajimehoshi@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 : hajimehoshi
  Commit : e80c576c7bdcd567ba96a4e1ceb9f845dcf1168e
  Date   : Fri May 26 05:49:36 2017
  Subject: Recode shared memory usage on Windows, macOS and NaCl

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:effective_size_avg/browse_news/browse_news_flipboard
  Change       : 206.02% | 527140933.167 -> 1613137138.0

Revision             Result                       N
chromium@474923      527140933 +- 41645083        6      good
chromium@474927      560212682 +- 32861872        6      good
chromium@474929      549064451 +- 21509400        6      good
chromium@474930      1615685873 +- 115334095      6      bad       <--
chromium@474936      1605640841 +- 159337625      6      bad
chromium@474949      1595210256 +- 145158341      6      bad
chromium@474975      1604726482 +- 110988487      6      bad
chromium@475027      1633162468 +- 80015828       6      bad
chromium@475131      1613137138 +- 100387499      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=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.news.flipboard 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/8975334844057153856


For feedback, file a bug with component Speed>Bisection

I guess this is related to https://bugs.chromium.org/p/chromium/issues/detail?id=713768. erikchen@, wdyt?

The CL in question claims that your patch added ~1GB to effective size on Windows...

Looking at the trace...this indeed appears to be the case.
https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_0-2017-06-30_13-46-59-61555.html

At a guess, there are two issues:
  1) https://bugs.chromium.org/p/chromium/issues/detail?id=713768 - resident vs. ritual
  2) ownership edges. I suspect they're not being correctly set up with cc, which has 432MB.

So I think we need to fix 713768 asap, right?

For 2), I'm not sure our fix for ownership edges is reflected on what you are looking at. Was the trace result at #19 taken with the latest chromium?

If the ETA for a fix for Issue# 713768 is not this week, we should revert e80c576c7bdcd567ba96a4e1ceb9f845dcf1168e

Hmm, 

Oops sorry, I sent this message accidentally...

> If the ETA for a fix for Issue# 713768 is not this week

Hmm, I don't think this happens until the end of this week.

My idea is to rename the current 'size' attribute to 'virtual_size' like DiscardableSharedMemory. If this is not a good idea, I'm OK to revert the CL (this just adds the hooks and I believe reverting will not cause crashes).

ssid@, primiano@, erikchen@, what do you think?

Wait we are confusing 2 different issues here.
The regression in UMA is definitely not related to the regression in all_process:effective_size in the graphs. The CL pointed just changed some accounting of shared memory in memory-infra.
process:effective_size should be removed from the metrics we alert on in telemetry. It shows a total number which does not mean anything in reality since we have different kinds of allocators and some do not report resident size anymore.
The number we record on UMA is different from the number the graphs in telemetry report. The UMA number is the total Private memory of all processes on Win. The graphs show a total number reported by all allocators and regression is because of shared memory total, which cannot be caught by UMA.

Having said that the regression on UMA could be a real regression. Do we see this regression on other platforms?

Looking at the latest graphs the regression has gone away:
https://uma.googleplex.com/timeline_v2?sid=d7224a89838dc99d00a6e1ec57dd1935

 Issue 737672  has been merged into this issue.

Based on the discussion in this bug, it looks like it was a reporting change or it was related to the blocking bug, so I'm closing this.

It looks like after this bug was filed though, there was a significant regression in the same metric, so I filed bug 752335 to investigate.