|
|
DescriptionEnable cc_perftests through chromium.perf.json.
BUG=392620
Patch Set 1 : #
Total comments: 2
Patch Set 2 : Removed GPU bots. #
Total comments: 2
Patch Set 3 : Removed --single-process-tests flag. #Messages
Total messages: 19 (6 generated)
simonhatch@chromium.org changed reviewers: + phajdan.jr@chromium.org
simonhatch@chromium.org changed reviewers: + danakj@chromium.org
Patchset #1 (id:1) has been deleted
danakj: Are these bots ok? I just enabled this on everything.
LGTM -gpu bots https://codereview.chromium.org/963203002/diff/20001/testing/buildbot/chromiu... File testing/buildbot/chromium.perf.json (right): https://codereview.chromium.org/963203002/diff/20001/testing/buildbot/chromiu... testing/buildbot/chromium.perf.json:74: "Win 7 ATI GPU Perf": { GPU isn't necessary, there are no GPU related tests.
https://codereview.chromium.org/963203002/diff/20001/testing/buildbot/chromiu... File testing/buildbot/chromium.perf.json (right): https://codereview.chromium.org/963203002/diff/20001/testing/buildbot/chromiu... testing/buildbot/chromium.perf.json:74: "Win 7 ATI GPU Perf": { On 2015/02/27 19:59:59, danakj wrote: > GPU isn't necessary, there are no GPU related tests. Done.
danakj@chromium.org changed reviewers: + enne@chromium.org
LGTM
The CQ bit was checked by sullivan@chromium.org
The CQ bit was unchecked by sullivan@chromium.org
On 2015/02/27 20:12:40, danakj wrote: > LGTM phajdan: Need owners approval, ptal
https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... File testing/buildbot/chromium.perf.json (right): https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... testing/buildbot/chromium.perf.json:7: "args": ["cc_perftests", "--single-process-tests"] Please don't pass this from command line (--single-process-tests). Instead the test binary should specify tests should run serially, see e.g. interactive_ui_tests or media_perftests.
https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... File testing/buildbot/chromium.perf.json (right): https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... testing/buildbot/chromium.perf.json:7: "args": ["cc_perftests", "--single-process-tests"] On 2015/03/02 18:52:25, Paweł Hajdan Jr. wrote: > Please don't pass this from command line (--single-process-tests). Instead the > test binary should specify tests should run serially, see e.g. > interactive_ui_tests or media_perftests. Ah I see they run serially anyway (crbug.com/236893). Added a stdio flag so that the output is still parsed/sent to dashboard.
On 2015/03/02 21:03:58, shatch wrote: > https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... > File testing/buildbot/chromium.perf.json (right): > > https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... > testing/buildbot/chromium.perf.json:7: "args": ["cc_perftests", > "--single-process-tests"] > On 2015/03/02 18:52:25, Paweł Hajdan Jr. wrote: > > Please don't pass this from command line (--single-process-tests). Instead the > > test binary should specify tests should run serially, see e.g. > > interactive_ui_tests or media_perftests. > > Ah I see they run serially anyway (crbug.com/236893). Added a stdio flag so that > the output is still parsed/sent to dashboard. Repeating my question on https://codereview.chromium.org/962863002/: Why is this stdio flag needed? Is this something devs would want to set when running the test too? If so, should ti just be on by default?
On 2015/03/02 22:59:18, Nico wrote: > On 2015/03/02 21:03:58, shatch wrote: > > > https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... > > File testing/buildbot/chromium.perf.json (right): > > > > > https://codereview.chromium.org/963203002/diff/40001/testing/buildbot/chromiu... > > testing/buildbot/chromium.perf.json:7: "args": ["cc_perftests", > > "--single-process-tests"] > > On 2015/03/02 18:52:25, Paweł Hajdan Jr. wrote: > > > Please don't pass this from command line (--single-process-tests). Instead > the > > > test binary should specify tests should run serially, see e.g. > > > interactive_ui_tests or media_perftests. > > > > Ah I see they run serially anyway (crbug.com/236893). Added a stdio flag so > that > > the output is still parsed/sent to dashboard. > > Repeating my question on https://codereview.chromium.org/962863002/: Why is this > stdio flag needed? Is this something devs would want to set when running the > test too? If so, should ti just be on by default? So my understanding is that there are only 2 supported paths to generating the data for perf dashboard upload. In runtest.py we eventually call _SendResultsToDashboard, which takes a log_processor. Telemetry tests always generate a TelemetryResultsProcessor which outputs the new Telemetry ChartJson format which is supported by the dashboard. Otherwise, the only log_processor with support for generating data for the dashboard right now is GraphingLogProcessor, which expects to parse the stdio of the test line by line. It's not something that devs need when running the test, only needed so that the results get sent to the perf dashboard.
Cool, thanks. Should this flag be in the "which tests to run" metadata file, or should it be in the bot config? The latter seems like the "more correct" place to me, but I'm not sure.
On 2015/03/03 20:04:09, Nico wrote: > Cool, thanks. > > Should this flag be in the "which tests to run" metadata file, or should it be > in the bot config? The latter seems like the "more correct" place to me, but I'm > not sure. Maybe the metadata file isn't the best spot, but another option is we could also add it to the gtest_perf_test.py script since I'm guessing all callers will want the stdio so that this is uploaded to the dashboard. I'm not super familiar with the whole build system, what do you mean by the bot config?
LGTM Please move away from scraping the log output, and use JSON format for this instead.
On 2015/03/03 20:10:36, shatch wrote: > On 2015/03/03 20:04:09, Nico wrote: > > Cool, thanks. > > > > Should this flag be in the "which tests to run" metadata file, or should it be > > in the bot config? The latter seems like the "more correct" place to me, but > I'm > > not sure. > > Maybe the metadata file isn't the best spot, but another option is we could also > add it to the gtest_perf_test.py script since I'm guessing all callers will want > the stdio so that this is uploaded to the dashboard. > > I'm not super familiar with the whole build system, what do you mean by the bot > config? There are lots of files that explain to the bots what build steps to run, which flags to pass for each step, which env vars to set, and whatnot. But if Pawel i shappy with this cl, go ahead as-is and ignore me :-) |