GitLab

Scheduling these benchmarks on the perf waterfall. Turns out, even if
you know a benchmark won't be run on certain types of devices, you still
have to add them to the sharding map for those configurations.
Example: loading.mobile is sharded onto windows bots even though the tests are
specifically for mobile configurations. We should change the unscheduled benchmarks
check of perf_data_generator.py in the future. perf_data_generator seems to do the
right thing in not scheduling the benchmarks on configurations that don't run them.

BUG=chromium:789571

Change-Id: I1a7b785cc04f5b207d6a60d05b8177bc9e7df37c
Reviewed-on: https://chromium-review.googlesource.com/801580Reviewed-by: Ned Nguyen <nednguyen@google.com>
Commit-Queue: Ashley Enstad <ashleymarie@chromium.org>
Cr-Commit-Position: refs/heads/master@{#520756}