All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=766889

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=0631f4064eafaea4a81025c75cfc11cb706514e5d5d2a7fc71aa1e8c168489a3


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

chromium-rel-win8-dual

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8967979977796362704

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : media_perftests
  Metric       : vector_math_fmul/unoptimized
  Change       : 1.54% | 724.453042354 -> 733.091628103

Suspected Commit Range
  2 commits in range
  https://chromium.googlesource.com/chromium/src/+log/27c7f330af36d90d1c9f8b504a02dd20f8a1a006..d0ce886d1bceac7bbc3a440254c071c9d6f1cf3c


Revision             Result                  N
chromium@502532      724.453 +- 18.7575      14       good
chromium@502533      ---                     ---      build failure
chromium@502534      728.39 +- 28.6823       14       bad
chromium@502536      734.359 +- 14.1753      6        bad
chromium@502540      733.013 +- 30.4478      14       bad
chromium@502548      733.092 +- 18.9477      9        bad

To Run This Test
  .\src\out\Release\media_perftests.exe --single-process-tests

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8967979977796362704


For feedback, file a bug with component Speed>Bisection

+ethannicholas, 

Bisect narrowed the regression in speed of calculation for test "	vector_math_fmul/unoptimized" to 2 commits. One of the commits is an unrelated change in layout test test expectations, and one of them is a skia roll containing three changes (https://chromium.googlesource.com/chromium/src/+/ae7428199aeaf4e66ed9b71cdac8e9617ae5507a) that you made. If you look at the graph (https://chromeperf.appspot.com/group_report?bug_id=766889) it looks like your change is returning this to an earlier state. Please take a look but feel free to mark as wontfix if this is expected.

Dale, I think you're the person who understands how "vector_math_fmul/unoptimized" test works. Is a 3.5% to 1.5% regression enough to worry about?

unoptimized is really just used as a baseline, so no need to monitor it for regressions. This seems some sort of linking/code folding related regression that doesn't really matter.

Thanks Dale. I filed  issue 767124  to figure out how to no longer alert on such things.