All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=778619

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=a2f2b720fd332fc9ebbc9081187caa2b5eac1a4671372a915857d503d53660d8


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

chromium-rel-win8-dual

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8964676865101300096

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

Hi brianosman@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 : Brian Osman
  Commit : 1cb41717bc4a44272eab48bd47ca7579425dc22e
  Date   : Mon Oct 23 15:27:12 2017
  Subject: Remove color space xform support from SkSL

Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_constant_full_page_raster_10000_pixels_per_second
  Change       : 27.11% | 2.10026696004 -> 2.66962162383

Revision                             Result                    N
chromium@510776                      2.10027 +- 0.251586       6      good
chromium@510813                      2.31416 +- 0.230902       6      good
chromium@510831                      2.29325 +- 0.150309       6      good
chromium@510833                      1.86586 +- 0.0724131      6      good
chromium@510833,skia@e25d119709      1.83358 +- 0.0678622      6      good
chromium@510833,skia@1cb41717bc      2.7092 +- 0.0688985       6      bad       <--
chromium@510834                      2.75883 +- 0.0771006      6      bad
chromium@510836                      2.70643 +- 0.0419774      6      bad
chromium@510840                      2.62386 +- 0.126601       6      bad
chromium@510849                      2.66962 +- 0.0944033      6      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=text.constant.full.page.raster.10000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8964676865101300096


For feedback, file a bug with component Speed>Bisection

I don't see how my CL could have caused a regression - it simply removed dead code (eliminated an unused feature from our shader compiler).

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8962932803736782976

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

Suspected Commit
  Author : Brian Osman
  Commit : 1cb41717bc4a44272eab48bd47ca7579425dc22e
  Date   : Mon Oct 23 15:27:12 2017
  Subject: Remove color space xform support from SkSL

Bisect Details
  Configuration: win_8_perf_bisect
  Benchmark    : thread_times.tough_scrolling_cases
  Metric       : thread_raster_cpu_time_per_frame/text_constant_full_page_raster_10000_pixels_per_second
  Change       : 27.79% | 2.10629102187 -> 2.69154459123

Revision                             Result                    N
chromium@510776                      2.10629 +- 0.0859047      6      good
chromium@510813                      2.26556 +- 0.159139       6      good
chromium@510831                      2.33011 +- 0.205036       6      good
chromium@510833                      1.85209 +- 0.0852292      6      good
chromium@510833,skia@e25d119709      1.85075 +- 0.0428668      6      good
chromium@510833,skia@1cb41717bc      2.74872 +- 0.0565         6      bad       <--
chromium@510834                      2.73131 +- 0.111641       6      bad
chromium@510836                      2.70307 +- 0.100307       6      bad
chromium@510840                      2.64056 +- 0.138006       6      bad
chromium@510849                      2.69154 +- 0.0515743      6      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=text.constant.full.page.raster.10000.pixels.per.second thread_times.tough_scrolling_cases

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8962932803736782976


For feedback, file a bug with component Speed>Bisection

both bisects point to brianosman's CL, but I'll re-kick on pinpoint to see if it gives us a clearer picture of what's happening, it looks like maybe this is bimodal?

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1198f29f040000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/149ff16f040000

The pinpoint jobs are pretty clearly pointing at the change to clang, right?

I'm not sure how to read the pinpoint jobs, but if it is the clang switch, I don't think we'll be going after this single regression.