In issue 921342 the baseline time unit was changed from nanoseconds to milliseconds. Additionally the scaling factor was inverted so we ended up with times that were in femtoseconds. Could we get the following scaling applied to the angle_perftests data:

1) 

bots: "Win 7 NVIDIA GPU Perf" and "win10-perf"
tests: everything generated by angle_perftests
revisions: 620237-present**
scaling: divide by 10e12

Applies to chromeperf v2spa dashboard.

**note: last revision must be after the fix in issue 921342

2) 

bots: "Win 7 NVIDIA GPU Perf" and "android-nexus5x-perf"
tests: everything generated by angle_perftests
revisions: 608300-620236
scaling: divide by 10e6

Only applies to chromeperf dashboard.

There are a few "ms" values in these tests that will be incorrectly scaled but I'm fine with that. We can file issues to have those adjusted individually if it comes up. But I don't expect it to be necessary as they are not very frequently used tests.
 

In issue 921342 the baseline time unit was changed from nanoseconds to milliseconds. Additionally the scaling factor was inverted so we ended up with times that were in femtoseconds. Could we get the following scaling applied to the angle_perftests data:

1) 

bots: "Win 7 NVIDIA GPU Perf" and "win10-perf"
tests: everything generated by angle_perftests
revisions: 620237-present**
scaling: divide by 10e12

Applies to chromeperf and v2spa dashboards.

**note: last revision must be after the fix in issue 921342

2) 

bots: "Win 7 NVIDIA GPU Perf" and "android-nexus5x-perf"
tests: everything generated by angle_perftests
revisions: 608300-620236
scaling: divide by 10e6

Only applies to chromeperf dashboard.

There are a few "ms" values in these tests that will be incorrectly scaled but I'm fine with that. We can file issues to have those adjusted individually if it comes up. But I don't expect it to be necessary as they are not very frequently used tests.