New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 758750 link

Starred by 1 user
Status: Fixed
Owner:
█ kbr@chromium.org
OOO until 2019-01-24
Closed: Aug 2017
Cc:
ynovikov@chromium.org
geoffl...@chromium.org
jmad...@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Bug

Blocked on:
issue 737280
Blocking:
issue 713196



Sign in to add a comment

Run more tests on Win AMD tryservers again

Project Member Reported by kbr@chromium.org, Aug 24 2017 
In Issue 737280 several test suites were disabled on the Windows AMD tryserver because of a lack of capacity. It turns out that 8 of the 30 bots had been accidentally removed from the Swarming pool, and this was probably a major contributing factor to the lack of capacity.

At this point the bots appear only lightly loaded:

https://viceroy.corp.google.com/chrome_infra/Jobs/pools?service_name=chromium-swarm&pool=.*gpu%3A1002%3A6613.*&duration=7d&refresh=-1

I think it is safe to add back most of the rest of the test suites on to this bot configuration and monitor the above graph for a few days to make sure we're not over capacity.
Project Member

Comment 1 by bugdroid1@chromium.org, Aug 25 2017 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ee458da4102bf149a5211a19485225ae329b1394

commit ee458da4102bf149a5211a19485225ae329b1394
Author: Kenneth Russell <kbr@chromium.org>
Date: Fri Aug 25 00:56:28 2017

Run webgl2_conformance_tests on AMD on win_angle_rel_ng.

There's extra capacity at this point and this will catch errors before
they reach the tree.

BUG= 758750 
TBR=ynovikov@chromium.org
NOTRY=true

Cq-Include-Trybots: master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: I739ee81b49c7e79150f06f17927acb70984ee5ac
Reviewed-on: https://chromium-review.googlesource.com/634517
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#497275}
[modify] https://crrev.com/ee458da4102bf149a5211a19485225ae329b1394/content/test/gpu/generate_buildbot_json.py
[modify] https://crrev.com/ee458da4102bf149a5211a19485225ae329b1394/testing/buildbot/chromium.gpu.fyi.json

Comment 2 by kbr@chromium.org, Aug 29 2017 

It looks like that might be the only test suite we can safely re-enable on these tryservers. Check the usage graph over the past couple of days:

https://viceroy.corp.google.com/chrome_infra/Jobs/pools?service_name=chromium-swarm&pool=.*gpu%3A1002%3A6613.*pool%3AChrome%24&duration=7d&refresh=-1

The pool is in full use with only a couple of tryjobs in the queue. It's not completely clear to me how many concurrent tryjobs would cause shards to start timing out.

Are there other test suites we should consider re-enabling at this point, or should we close this as fixed?

Comment 3 by ynovikov@chromium.org, Aug 30 2017 

I think we might have some capacity to reenable angle_deqp_gles3_d3d11_tests if issue 713196 will be fixed.

This issue can be closed, I think.

Comment 4 by kbr@chromium.org, Aug 30 2017

Blocking: 713196
Status: Fixed (was: Assigned)
With the recent split to put the dEQP tests on their own bots, we may be able to turn those tests on again on the dEQP trybots.

Thanks for the feedback. Closing this as fixed.
►

Sign in to add a comment