All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=727466

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDghqaTtQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDghs3RuQoM


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

chromium-rel-win10
chromium-rel-win7-gpu-ati

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8978200790268178656

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

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : smoothness.top_25_smooth
  Metric       : first_gesture_scroll_update_latency/Wikipedia (1 tab)

Revision             Result                  N
chromium@475131      8.51276 +- 27.5275      21      good
chromium@475225      10.785 +- 23.7434       21      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=Wikipedia..1.tab. smoothness.top_25_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8978200790268178656

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6440635797602304


| 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/8978193422152795968

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

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : smoothness.top_25_smooth
  Metric       : first_gesture_scroll_update_latency/Wikipedia (1 tab)

Revision             Result                  N
chromium@475131      11.2776 +- 22.1898      21      good
chromium@475225      10.9764 +- 22.3426      21      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=Wikipedia..1.tab. smoothness.top_25_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8978193422152795968

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6440635797602304


| 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!

assigned the test owner.

Hi vmiura@,

The bisect results say "11.2776 +- 22.1898". Does this means that noise range is wider than actual result value? If so, it seems this test does not work to catch real regressions.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8976996587343545152

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

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : smoothness.top_25_smooth
  Metric       : first_gesture_scroll_update_latency/Wikipedia (1 tab)

Revision             Result                  N
chromium@475100      8.19695 +- 27.736       21      good
chromium@475250      9.86481 +- 25.9688      21      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=Wikipedia..1.tab. smoothness.top_25_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976996587343545152

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5036379362295808


| 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/8976989243878157696

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

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : smoothness.top_25_smooth
  Metric       : first_gesture_scroll_update_latency/Wikipedia (1 tab)

Revision             Result                  N
chromium@475100      10.5935 +- 23.3959      21      good
chromium@475250      9.72581 +- 26.1771      21      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=Wikipedia..1.tab. smoothness.top_25_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976989243878157696

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5036379362295808


| 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/8975658352410039664

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

Suspected Commit
  Author : eseckler
  Commit : e356c64f253923983248656d304d8aa471aad5f0
  Date   : Fri May 26 12:59:36 2017
  Subject: [cc] Plumb BeginFrameAcks through SurfaceManager to DisplayScheduler.

Bisect Details
  Configuration: winx64_10_perf_bisect
  Benchmark    : smoothness.top_25_smooth
  Metric       : first_gesture_scroll_update_latency/Wikipedia (1 tab)
  Change       : 302.32% | 2.26804761905 -> 9.1249047619

Revision             Result                   N
chromium@474708      2.26805 +- 11.4697       21      good
chromium@474967      1.71571 +- 0.405461      14      good
chromium@474983      2.592 +- 11.2645         14      good
chromium@474987      1.7765 +- 0.684394       14      good
chromium@474988      1.75321 +- 0.482326      14      good
chromium@474989      10.4451 +- 25.5668       21      bad       <--
chromium@474992      10.4399 +- 20.4945       14      bad
chromium@474999      11.6932 +- 15.369        14      bad
chromium@475031      8.44878 +- 15.5216       9       bad
chromium@475097      9.84343 +- 20.113        14      bad
chromium@475225      9.1249 +- 26.8086        21      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=Wikipedia..1.tab. smoothness.top_25_smooth

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975658352410039664

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5845286036963328


| 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!

I'm deduping this issue again, since this is very likely caused by a different issue (though by the same patch) than other regressions in  bug 727513 .

It seems that, sometimes, the DisplayScheduler triggers an early deadline for a proactive BeginFrame without updates, which means that the renderer's later frame submission is postponed into the next frame.

This issue has been fixed in https://chromium-review.googlesource.com/c/536932/ very recently, and since then the metric seems to have returned to original levels. I'll keep an eye out if it stays there, since this was a flaky regression.

Metric has reverted to and remained at original levels, so closing this.