[Pywebsocket PerformanceTests 1/2] Add PerformanceTests/Pywebsocket

[Pywebsocket PerformanceTests 1/2] Add PerformanceTests/Pywebsocket

This CL adds PerformanceTests test scripts that will be used by [2/2].
These tests measure execution time of sending/receiving data of 10MB each
by XHR, WebSocket, and fetch.

WebSocket send benchmarks are disabled because there is a performance
bottleneck in pywebsocket side when running without native modules.

[1/2] PerformanceTests scripts: This CL
[2/2] Telemetry: https://codereview.chromium.org/736653002/

BUG=432408
Committed: https://crrev.com/5e3da089ef915c6870dc3db8c220841563567a82
git-svn-id: svn://svn.chromium.org/blink/trunk@202031 bbb929c8-8fbe-4397-9dbb-9b2b20218538

[hiroshige, 2014-11-18 10:33:07]

Patchset #1 (id:1) has been deleted

[hiroshige, 2015-04-08 08:02:13]

hiroshige@chromium.org changed reviewers:
+ tyoshino@chromium.org, yhirano@chromium.org

[hiroshige, 2015-04-08 08:52:03]

Patchset #4 (id:80001) has been deleted

[hiroshige, 2015-04-08 08:52:14]

Patchset #4 (id:100001) has been deleted

[tyoshino (SeeGerritForStatus), 2015-05-26 08:12:57]

Is this ready for review?

[hiroshige, 2015-05-26 08:22:28]

On 2015/05/26 08:12:57, tyoshino wrote:
> Is this ready for review?

Not yet. (Still debugging on Windows)

[hiroshige, 2015-08-18 04:57:22]

Now this CL is ready for review, PTAL.

Currently I put many configurations for testing.
Should we select a subset of them to commit and run on perfbots?

[yhirano, 2015-08-18 12:50:40]

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/generate.py (right):

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/generate.py:47: test_file_content =
(test_file_content_template %
I would prefer |format| function because we can use keyword parameters.

'{hoge}'.format(hoge='fuga')

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/generate.py:80: # Disable
XHR-send-*-window-sync tests.
receive?

[hiroshige, 2015-08-19 08:58:28]

Patchset #7 (id:180001) has been deleted

[hiroshige, 2015-08-19 08:58:53]

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/generate.py (right):

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/generate.py:47: test_file_content =
(test_file_content_template %
On 2015/08/18 12:50:40, yhirano wrote:
> I would prefer |format| function because we can use keyword parameters.
> 
> '{hoge}'.format(hoge='fuga')

Done.

https://codereview.chromium.org/738753002/diff/160001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/generate.py:80: # Disable
XHR-send-*-window-sync tests.
On 2015/08/18 12:50:40, yhirano wrote:
> receive?

Done.

[yhirano, 2015-08-20 06:17:49]

https://codereview.chromium.org/738753002/diff/200001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/util_performance_test.js (right):

https://codereview.chromium.org/738753002/diff/200001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/util_performance_test.js:54: done:
function() { doAction(config, isWorker, 'stop'); },
This may be not a problem of this CL, but doAction changes config and
worker.onMessage. Calling it more than once looks dangerous.
How about setting worker.onmessage in a separate function (initWorker for
example)?

[hiroshige, 2015-09-07 07:29:55]

https://codereview.chromium.org/738753002/diff/200001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/util_performance_test.js (right):

https://codereview.chromium.org/738753002/diff/200001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/util_performance_test.js:54: done:
function() { doAction(config, isWorker, 'stop'); },
On 2015/08/20 06:17:49, yhirano wrote:
> This may be not a problem of this CL, but doAction changes config and
> worker.onMessage. Calling it more than once looks dangerous.
> How about setting worker.onmessage in a separate function (initWorker for
> example)?

Done, created separate |config| for each doAction() call.
Yes, pywebsocket side should be refactored (and started refactoring but will
take some time and I don't want the refactoring to block this CL).

[yhirano, 2015-09-07 11:35:37]

lgtm

[hiroshige, 2015-09-08 05:51:30]

The CQ bit was checked by hiroshige@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-09-08 05:51:37]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/738753002/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/738753002/220001

[tyoshino (SeeGerritForStatus), 2015-09-08 06:22:36]

lgtm

https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/util_performance_test.js (right):

https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/util_performance_test.js:30:
numIterations: numIterations + numWarmUpIterations + 1,
please add a comment about +1

[hiroshige, 2015-09-10 05:36:44]

https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
File PerformanceTests/Pywebsocket/resources/util_performance_test.js (right):

https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
PerformanceTests/Pywebsocket/resources/util_performance_test.js:30:
numIterations: numIterations + numWarmUpIterations + 1,
On 2015/09/08 06:22:36, tyoshino wrote:
> please add a comment about +1

Done.

[hiroshige, 2015-09-10 05:37:18]

The CQ bit was checked by hiroshige@chromium.org

[hiroshige, 2015-09-10 05:37:19]

The patchset sent to the CQ was uploaded after l-g-t-m from
tyoshino@chromium.org, yhirano@chromium.org
Link to the patchset: https://codereview.chromium.org/738753002/#ps240001
(title: "Comment for +1.")

[commit-bot: I haz the power, 2015-09-10 05:37:33]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/738753002/240001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/738753002/240001

[tyoshino (SeeGerritForStatus), 2015-09-10 05:42:09]

On 2015/09/10 05:36:44, hiroshige wrote:
>
https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
> File PerformanceTests/Pywebsocket/resources/util_performance_test.js (right):
> 
>
https://codereview.chromium.org/738753002/diff/220001/PerformanceTests/Pywebs...
> PerformanceTests/Pywebsocket/resources/util_performance_test.js:30:
> numIterations: numIterations + numWarmUpIterations + 1,
> On 2015/09/08 06:22:36, tyoshino wrote:
> > please add a comment about +1
> 
> Done.

Thanks. lgtm

[commit-bot: I haz the power, 2015-09-10 06:22:35]

Committed patchset #9 (id:240001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=202031

[commit-bot: I haz the power, 2015-09-23 12:07:57]

Patchset 9 (id:??) landed as
https://crrev.com/5e3da089ef915c6870dc3db8c220841563567a82