Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
Committed: https://crrev.com/f553491c9104f998b48f50c66da1b14473af053d
Cr-Commit-Position: refs/heads/master@{#418508}

[hiroshige, 2016-09-09 13:00:11]

Description was changed from

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

BUG=645363
==========

to

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

[hiroshige, 2016-09-09 13:00:15]

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

[commit-bot: I haz the power, 2016-09-09 13:00:26]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[hiroshige, 2016-09-09 13:22:34]

Description was changed from

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

to

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

[hiroshige, 2016-09-09 13:29:17]

hiroshige@chromium.org changed reviewers:
+ fs@opera.com, kouhei@chromium.org, yhirano@chromium.org

[hiroshige, 2016-09-09 13:29:18]

PTAL.

yhirano@, please check the CL description; I expect the performance regression
in this case (a kind of microbenchmark of sync XHR) is safe to ignore unless it
hits real use cases, but please double check.

kouhei@ or fs@, PTAL as the owners of blink_perf.svg performance tests.

[commit-bot: I haz the power, 2016-09-09 15:34:49]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-09 15:34:50]

Dry run: This issue passed the CQ dry run.

[yhirano, 2016-09-12 10:37:16]

lgtm

[fs, 2016-09-12 10:58:01]

On 2016/09/12 at 10:37:16, yhirano wrote:
> lgtm

Sorry for the delay, this look reasonable to me. LGTM.

I must admit though that I feel somewhat alarmed by this "neutering" of the
MemoryCache (now sans the "Cache" essentially), and that the second part of the
statement: "This is expected and we can ignore the regression unless it causes
regressions in real use cases." could lead to a long tail of issues that may not
be obvious.

[hiroshige, 2016-09-14 07:00:12]

On 2016/09/12 10:58:01, fs wrote:
> On 2016/09/12 at 10:37:16, yhirano wrote:
> > lgtm
> 
> Sorry for the delay, this look reasonable to me. LGTM.
> 
> I must admit though that I feel somewhat alarmed by this "neutering" of the
> MemoryCache (now sans the "Cache" essentially), and that the second part of
the
> statement: "This is expected and we can ignore the regression unless it causes
> regressions in real use cases." could lead to a long tail of issues that may
not
> be obvious.

Yeah, but WeakMemoryCache [1] affects the lifetime of Resource largely, there
can be many microbenchmarks affected by [1], and I don't know which regressions
in such microbenchmarks are meaningful or not.

[hiroshige, 2016-09-14 07:00:19]

The CQ bit was checked by hiroshige@chromium.org

[commit-bot: I haz the power, 2016-09-14 07:00:34]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-14 08:00:11]

Description was changed from

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

to

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

[commit-bot: I haz the power, 2016-09-14 08:00:15]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-09-14 08:01:48]

Description was changed from

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363
==========

to

==========
Load the test page before measurement in PerfTestRunner.measurePageLoadTime()

Previously, loadFile() is called to load the test page during performance
measurement in PerfTestRunner.measurePageLoadTime(), and thus the results of
blink_perf.svg tests were affected by the performance of loadFile() (that is not
related to SVG).
This CL makes the loadFile() to be executed before measurement and thus avoids
blink_perf.svg regressions caused by loadFile() performance regression.

The loadFile() performance regression is caused by [1] because, after [1],
synchronous XHRs are evicted from MemoryCache when Oilpan GC is executed.
[1] https://codereview.chromium.org/2296913003
This is expected and we can ignore the regression unless it causes regressions
in real use cases.

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:linux_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq;master.tryserver.chromium.perf:winx64_10_perf_cq
BUG=645363

Committed: https://crrev.com/f553491c9104f998b48f50c66da1b14473af053d
Cr-Commit-Position: refs/heads/master@{#418508}
==========

[commit-bot: I haz the power, 2016-09-14 08:01:49]

Patchset 1 (id:??) landed as
https://crrev.com/f553491c9104f998b48f50c66da1b14473af053d
Cr-Commit-Position: refs/heads/master@{#418508}