All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=734720

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDg9tSPkwkM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgtquuugkM


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

chromium-rel-mac-retina
linux-release

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8976329630134507264

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

Bisect Details
  Configuration: linux_perf_bisect
  Benchmark    : smoothness.gpu_rasterization.tough_filters_cases
  Metric       : frame_times/IE PirateMark

Revision             Result                   N
chromium@479028      16.7221 +- 0.268708      21      good
chromium@479215      16.6998 +- 0.180564      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=IE.PirateMark smoothness.gpu_rasterization.tough_filters_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976329630134507264

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6440764881502208


| 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/8976324861223369424

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8976323500802488336

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

Bisect Details
  Configuration: linux_perf_bisect
  Benchmark    : smoothness.gpu_rasterization.tough_filters_cases
  Metric       : frame_times/IE PirateMark

Revision             Result                   N
chromium@479028      16.6962 +- 0.306452      21      good
chromium@479215      16.7088 +- 0.576505      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=IE.PirateMark smoothness.gpu_rasterization.tough_filters_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976324861223369424

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6440764881502208


| 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!

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

Bisect Details
  Configuration: linux_perf_bisect
  Benchmark    : smoothness.gpu_rasterization.tough_filters_cases
  Metric       : frame_times/Filter Terrain SVG

Revision             Result                   N
chromium@479028      16.8702 +- 0.878185      21      good
chromium@479215      16.9303 +- 0.965001      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=Filter.Terrain.SVG smoothness.gpu_rasterization.tough_filters_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976323500802488336

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6460984513789952


| 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!

Closing since the confidence intervals are wide and bisect cannot reproduce.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8976317313985567808

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

Bisect Details
  Configuration: linux_perf_bisect
  Benchmark    : smoothness.gpu_rasterization.tough_filters_cases
  Metric       : frame_times/Filter Terrain SVG

Revision             Result                   N
chromium@479028      16.7614 +- 0.494278      21      good
chromium@479215      16.7865 +- 0.579249      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=Filter.Terrain.SVG smoothness.gpu_rasterization.tough_filters_cases

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8976317313985567808

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=6460984513789952


| 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!

Looking just at linux-release it looks like the Ref version of Filter Terrain (and its non-named predecessor, http___letmespellitoutforyou.com_samples_svg_filter_terrain.svg), actually regressed first, back in 477411 - 477510.

Then the mainline (non-ref) version regressed in 479029 - 479215. Very strange.

Both ref and non-ref previously had isolated spikes up to the current level (30ms), about 3 spikes in ~9000 revs, but now those levels are sustained.

There are now intermittent spikes up to ~100ms.

Re #11, see https://chromeperf.appspot.com/report?sid=a8154347e4cae26b095fbf3081d4887398388b3802aed4cdcb7bed58c39bf7b0