All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=763896

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=d2b31809ceaa6a26b25cff538097132644f216d7b88fc90c9fac2a6d0932fcb8


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

chromium-rel-win10

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968746915154011904

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

Hi vmpstr@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 : Vladimir Levin
  Commit : c9aa97407de3305265e662d3dee32c2a006923b2
  Date   : Thu Sep 07 21:24:57 2017
  Subject: images: Plumb decoding to scale for JPEGs in software.

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:skia:effective_size_avg/load_news/load_news_qq
  Change       : 1152.01% | 403684.666667 -> 5054167.33333

Revision             Result                  N
chromium@500368      403685 +- 60222.5       6      good
chromium@500379      390225 +- 33241.8       6      good
chromium@500382      401450 +- 66878.0       6      good
chromium@500383      400608 +- 44691.3       6      good
chromium@500384      5062909 +- 48670.7      6      bad       <--
chromium@500385      5066925 +- 60167.7      6      bad
chromium@500390      4483291 +- 3179834      6      bad
chromium@500411      5083501 +- 56166.1      6      bad
chromium@500453      5054167 +- 34553.3      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=load.news.qq 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/8968746915154011904


For feedback, file a bug with component Speed>Bisection

 Issue 763922  has been merged into this issue.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968715885968181920

=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Vladimir Levin
  Commit : c9aa97407de3305265e662d3dee32c2a006923b2
  Date   : Thu Sep 07 21:24:57 2017
  Subject: images: Plumb decoding to scale for JPEGs in software.

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:skia:effective_size_avg/load_news/load_news_qq
  Change       : 1109.55% | 405072.0 -> 4899542.66667

Revision             Result                  N
chromium@500368      405072 +- 53331.2       6      good
chromium@500379      397887 +- 81080.7       6      good
chromium@500382      400042 +- 41914.9       6      good
chromium@500383      409144 +- 66797.4       6      good
chromium@500384      5051130 +- 49838.6      6      bad       <--
chromium@500385      4923579 +- 784513       6      bad
chromium@500390      3760874 +- 4551730      6      bad
chromium@500411      5073941 +- 70480.6      6      bad
chromium@500453      4899543 +- 961487       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=load.news.qq 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/8968715885968181920


For feedback, file a bug with component Speed>Bisection

I believe this to be a timing issue. See:
https://chromeperf.appspot.com/report?sid=ecfda8d3790fa7c9ab0361042dbc32d03934e87a48fdc7fb99b9261ce0e6bce4&start_rev=490644&end_rev=500919

There is a exact corresponding decrease in unaccounted discardable memory as there is an increase in skia cache memory. Overall, the chrome effective size remains roughly the same.

The reason I suspect that this is a timing issue is that I can reproduce this locally. However, I added a few debugging printfs to ensure that skia was caching things that it was supposed to be caching and I found that I could no longer reproduce the issue...

Overall, the graphs agree that in terms of overall memory this patch didn't seem to change anything.

I'm going to confirm with the rest of the graphs if they tell the same story. 

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968701939670548080

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968701792090440944

I started a few more bisects, since v8 that shows up here is highly unlikely to be related.

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

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

Revision             Result                     N
chromium@500271      117387858 +- 21156883      21      good
chromium@500308      116173671 +- 21631305      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=release_x64 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=load.news.nytimes 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/8968701792090440944


For feedback, file a bug with component Speed>Bisection

=== Auto-CCing suspected CL author jupvfranco@google.com ===

Hi jupvfranco@google.com, 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 : Juliana Franco
  Commit : f82ae284fa2b1724e3dc2df7f4dd8ffc819800c2
  Date   : Thu Sep 07 12:36:49 2017
  Subject: Remove the next field from JS functions.

Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:v8:heap:allocated_objects_size_avg/load_media/load_media_9gag
  Change       : 2.83% | 23795408.6667 -> 24468020.6667

Revision                           Result                  N
chromium@500421                    23795409 +- 385884      6      good
chromium@500424                    23789263 +- 397008      6      good
chromium@500426                    23789784 +- 384198      6      good
chromium@500426,v8@7b53a0e010      23790420 +- 239683      6      good
chromium@500426,v8@f82ae284fa      24333195 +- 425323      6      bad       <--
chromium@500426,v8@e67420cbc2      24293152 +- 506757      6      bad
chromium@500426,v8@357269a5d9      24277469 +- 348997      6      bad
chromium@500427                    24355891 +- 230014      6      bad
chromium@500432                    24287668 +- 673691      6      bad
chromium@500442                    24345253 +- 229229      6      bad
chromium@500463                    24386879 +- 244607      6      bad
chromium@500504                    24468021 +- 716233      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 --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=load.media.9gag 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/8968701939670548080


For feedback, file a bug with component Speed>Bisection

As per #8, the skia/discardable memory seems to be an accounting issue (skia goes up by exactly the same amount as discardable goes down); this seems to be due to timing, since I get varying results depending on the amount of logging that I add. (FWIW skia memory is coming via canvas rendering images which does a blink-side decode). 

For the v8 regressions, the bisect seemed to have found a possible suspect. 

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8968469241766237536

=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : Juliana Franco
  Commit : f82ae284fa2b1724e3dc2df7f4dd8ffc819800c2
  Date   : Thu Sep 07 12:36:49 2017
  Subject: Remove the next field from JS functions.

Bisect Details
  Configuration: win_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:v8:heap:allocated_objects_size_avg/load_media/load_media_9gag
  Change       : 1.85% | 23976660.2857 -> 24420602.2857

Revision                           Result                   N
chromium@500421                    23976660 +- 1579369      14      good
chromium@500426                    23914233 +- 1135076      14      good
chromium@500426,v8@7b53a0e010      23898289 +- 775575       9       good
chromium@500426,v8@f82ae284fa      24300782 +- 401394       9       bad       <--
chromium@500426,v8@e67420cbc2      24326177 +- 291891       9       bad
chromium@500426,v8@357269a5d9      24292234 +- 392967       9       bad
chromium@500427                    24239943 +- 826991       14      bad
chromium@500428                    24243106 +- 506619       9       bad
chromium@500429                    24346566 +- 422963       9       bad
chromium@500431                    24319879 +- 702935       21      bad
chromium@500441                    24344759 +- 755311       14      bad
chromium@500463                    24304815 +- 722456       21      bad
chromium@500504                    24420602 +- 732071       14      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=load.media.9gag 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/8968469241766237536


For feedback, file a bug with component Speed>Bisection

 Issue 765238  has been merged into this issue.

 Issue 765238  has been merged into this issue.

 Issue 765238  has been merged into this issue.

 Issue 765238  has been merged into this issue.

This needs investigation on the memory side, as there's nothing that Juliana could do about it (seems to be some GC heuristics, whatever).

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8967099680291509968

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

Bisect Details
  Configuration: mac_10_12_perf_bisect
  Benchmark    : system_health.memory_desktop
  Metric       : memory:chrome:all_processes:reported_by_chrome:cc:effective_size_avg/browse_media/browse_media_flickr_infinite_scroll

Revision             Result                    N
chromium@500342      329152505 +- 7375553      21      good
chromium@500435      328213344 +- 7263708      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=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.flickr.infinite.scroll 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/8967099680291509968


For feedback, file a bug with component Speed>Bisection

The graph for memory:chrome:renderer_processes:reported_by_chrome:v8:heap:allocated_objects_size_avg / load_media / load_media_9gag has recovered.

Other regressions are in skia and are not actionable per comment #8.

Closing.