New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 734700 link

Starred by 3 users
Status: WontFix
Owner:
mek@chromium.org
Closed: Jul 2017
Cc:
kouhei@chromium.org
█ nednguyen@chromium.org
ksakamoto@chromium.org
charliea@chromium.org
chiniforooshan@chromium.org
mek@chromium.org
█ kraynov@chromium.org
lanwei@chromium.org
█ nasko@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

6%-29.2% regression in loading.desktop at 479722:479886

Project Member Reported by chiniforooshan@chromium.org, Jun 19 2017 
See the link to graphs below.
Project Member

Comment 4 by 42576172...@developer.gserviceaccount.com, Jun 19 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstMeaningfulPaint_avg/pcv1-warm/FIFA

Revision             Result                  N
chromium@479749      505.185 +- 170.744      21      good
chromium@479883      503.903 +- 179.147      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=FIFA loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976333029817100608

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5820001765818368


| 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!
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Jun 19 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : page_cycler_v2_site_isolation.basic_oopif
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/http___www.fifa.com_

Revision             Result                  N
chromium@479749      150.216 +- 36.9431      21      good
chromium@479883      148.846 +- 24.9114      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=http...www.fifa.com. page_cycler_v2_site_isolation.basic_oopif

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976332939710588688

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5846243428794368


| 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!
Project Member

Comment 9 by 42576172...@developer.gserviceaccount.com, Jun 19 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstMeaningfulPaint_avg/pcv1-warm/FIFA

Revision             Result                  N
chromium@479749      509.824 +- 207.886      21      good
chromium@479883      506.492 +- 164.319      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=FIFA loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976324862974954624

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5820001765818368


| 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!
Project Member

Comment 10 by 42576172...@developer.gserviceaccount.com, Jun 19 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : page_cycler_v2_site_isolation.basic_oopif
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/http___www.fifa.com_

Revision             Result                  N
chromium@479749      149.545 +- 23.0477      21      good
chromium@479883      148.183 +- 9.56003      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=http...www.fifa.com. page_cycler_v2_site_isolation.basic_oopif

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976321085117700704

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5846243428794368


| 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!
Project Member

Comment 11 by 42576172...@developer.gserviceaccount.com, Jun 20 2017

Cc: mek@chromium.org
Owner: mek@chromium.org

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

Hi mek@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 : Marijn Kruisselbrink
  Commit : 469805742123ec65bdeea766d8960c5abd7f82e8
  Date   : Thu Jun 15 19:33:41 2017
  Subject: If the file to rename doesn't exist, return OK immediately.

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : loading.desktop
  Metric       : cpuTimeToFirstMeaningfulPaint_avg/pcv1-cold/AirBnB
  Change       : 5.16% | 1495.32755556 -> 1572.48122222

Revision             Result                  N
chromium@479749      1495.33 +- 137.537      9      good
chromium@479783      1493.33 +- 77.4792      6      good
chromium@479784      1492.98 +- 72.3925      6      good
chromium@479785      1606.15 +- 67.9008      6      bad       <--
chromium@479786      1577.25 +- 50.8916      6      bad
chromium@479788      1581.39 +- 55.2132      6      bad
chromium@479792      1600.85 +- 97.1923      6      bad
chromium@479800      1595.11 +- 173.076      9      bad
chromium@479816      1577.95 +- 126.416      9      bad
chromium@479883      1572.48 +- 112.52       9      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=AirBnB loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976324073615697008

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5059898267664384


| 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!

Comment 12 by benhenry@chromium.org, Jun 20 2017

Status: Assigned (was: Untriaged)

Comment 13 by mek@chromium.org, Jun 20 2017 

That's weird... I wouldn't expect a single FileExists check to cause any meaningful performance regression (and this CL did fix a bunch of performance regressions on many other bots...). If it really was that CL that caused this regression there isn't really anything I can do, and we'll probably just have to live with it. Maybe some combination of scheduling changes and optimizations for in-process mojo calls that are both being worked on could have some effect here (hopefully for the better), but otherwise not sure what's going on, and no idea why this change would cause any kind of perf regression...

Comment 14 by chiniforooshan@chromium.org, Jun 20 2017 

Bisect results do look noisy. Let's see if it points to the same CL if I run it again on a different page...
Project Member

Comment 16 by 42576172...@developer.gserviceaccount.com, Jun 20 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : loading.desktop
  Metric       : cpuTimeToFirstMeaningfulPaint_avg/pcv1-warm/2ch

Revision             Result                  N
chromium@479749      49.4942 +- 8.2184       21      good
chromium@479883      51.0286 +- 16.4181      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=2ch loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976236856053634720

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6118399987417088


| 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!
Project Member

Comment 18 by 42576172...@developer.gserviceaccount.com, Jun 20 2017 


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

Bisect Details
  Configuration: mac_air_perf_bisect
  Benchmark    : loading.desktop
  Metric       : cpuTimeToFirstMeaningfulPaint_avg/pcv1-warm/2ch

Revision             Result                  N
chromium@479749      49.0643 +- 5.90379      21      good
chromium@479883      48.5908 +- 4.43826      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=2ch loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976230495067006096

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6118399987417088


| 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!
Project Member

Comment 19 by 42576172...@developer.gserviceaccount.com, Jun 22 2017

Cc: kraynov@chromium.org
 Issue 735929  has been merged into this issue.
Project Member

Comment 20 by 42576172...@developer.gserviceaccount.com, Jun 22 2017 

 Issue 735929  has been merged into this issue.
Project Member

Comment 22 by 42576172...@developer.gserviceaccount.com, Jun 22 2017

Cc: lanwei@chromium.org
 Issue 735675  has been merged into this issue.

Comment 23 by mek@chromium.org, Jun 22 2017 

FWIW the memory "regression" of  issue 735675  seems to more or less match perfectly a memory "improvement" of a build that included r478720 (which had a bug that the CL being blamed here fixed). Similarly with 735929, where there also is a matching improvement in the build that includes r478720.

But some of the other graphs (like the original one in this bug) don't have a matching improvement in the build with r478720, so not sure what is going on there. On the other hand, while for example there definitely seems to be a clear regression in ChromiumPerf/chromium-rel-mac11-air/loading.desktop / cpuTimeToFirstMeaningfulPaint_avg / pcv1-cold / AirBnB, looking at the graph it seems like that regression at least partially already happened in the a few builds before the range that included my CL. So not sure what's going on.
Project Member

Comment 24 by 42576172...@developer.gserviceaccount.com, Jun 22 2017 


=== BISECT JOB RESULTS ===
Bisect failed unexpectedly

Bisect was aborted with the following:
  Step('Check bisect finished on revision chromium@479877,skia@2bd381bffd.fetch file 2bd381bffd36cdcffacf606d952547ce66fed7e9:DEPS') failed with return_code 1


Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : page_cycler_v2_site_isolation.basic_oopif
  Metric       : timeToFirstMeaningfulPaint_avg/pcv1-warm/http___www.fifa.com_

Revision                             Result                  N
chromium@479749                      455.067 +- 168.439      14      good
chromium@479817                      475.889 +- 15.4077      6       good
chromium@479850                      450.83 +- 192.844       14      good
chromium@479867                      452.992 +- 257.574      21      good
chromium@479877                      462.965 +- 175.673      14      good
chromium@479877,skia@643fcfa02f      451.732 +- 176.316      14      good
chromium@479877,skia@2bd381bffd      466.825 +- 216.561      21      good
chromium@479878                      491.633 +- 40.9896      14      bad
chromium@479880                      477.54 +- 80.8417       14      bad
chromium@479883                      486.543 +- 13.3841      14      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=http...www.fifa.com. page_cycler_v2_site_isolation.basic_oopif

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976074727953861648

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4640840774844416


| 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!
Project Member

Comment 26 by 42576172...@developer.gserviceaccount.com, Jun 23 2017 


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

Bisect Details
  Configuration: mac_10_11_perf_bisect
  Benchmark    : page_cycler_v2_site_isolation.basic_oopif
  Metric       : timeToFirstMeaningfulPaint_avg/pcv1-warm/http___www.fifa.com_

Revision             Result                  N
chromium@479749      452.413 +- 229.73       21      good
chromium@479883      468.506 +- 160.142      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=http...www.fifa.com. page_cycler_v2_site_isolation.basic_oopif

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976045525875909360

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4640840774844416


| 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!

Comment 27 by mek@chromium.org, Jul 10 2017

Status: WontFix (was: Assigned)
Marking this as WontFix, as it's not really clear if there is actually something going wrong here, and even if my CL is actually to blame that just means that having working localstorage makes loading slower than when localstorage is potentially broken. Some of the same-process mojo optimizations might make some improvement in this area, but either way there is nothing actionable to do on this bug.

Comment 28 by mek@chromium.org, Jul 10 2017

Components: Blink>Storage>DOMStorage
►

Sign in to add a comment