All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=760173

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=adff91f5e81e6b9d146b2f7a0f58e721af11602f6b5ac5e342419e61e6a7c6fb


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

android-nexus5
android-nexus5X
android-one
android-webview-nexus6

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969912168141940128

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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:all_processes:reported_by_chrome:skia:effective_size_avg/browse_news/browse_news_cnn

Revision             Result                 N
chromium@496767      2088455 +- 196971      21      good
chromium@497657      2091314 +- 185349      21      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-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.news.cnn 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/8969912168141940128


For feedback, file a bug with component Speed>Bisection

Juan, can you help triage?

Especially for the timeseries that was bisected, it seems like a clear 1mib regression in skia, which shows up in the traces like this:
Before: https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_9-2017-08-23_22-01-30-29775.html
After: https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_9-2017-08-26_09-52-47-89886.html

When I click skia/renderer memory, the trace shows 900kib difference, made up of CNN Sans fonts (screenshots attached). Could this be a timing issue with the fonts loaded? Otherwise, why doesn't bisect repro at all?

Deleted: skia_before.png 39.1 KB

	skia_before.png 39.1 KB View Download

Deleted: skia_after.png 113 KB

	skia_after.png 113 KB View Download

Just recording what I've checked so far:

- the bisect reproduced no changes in the distribution before/after, so I don't think this is just a timing issue (the change on the perf dashboard is pretty clear, not noisy)

- there were no obvious device changes

will keep looking ...

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964840392740404064

Still puzzled, but I have a theory. Kicked off a new bisect with: (1) a slightly larger range and (2) running a few more stories.

Maybe there is some strange interaction going on between multiple stories?

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

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:all_processes:reported_by_chrome:skia:effective_size_avg/browse_news/browse_news_cnn

Revision             Result                 N
chromium@496601      2096047 +- 208585      21      good
chromium@497657      2075201 +- 168385      20      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-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.news 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/8964840392740404064


For feedback, file a bug with component Speed>Bisection

Juan, any update here?

Nope, and the issue is sufficiently old :(