Add paint testing mode subsequence_caching_disabled

Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect

Committed: https://crrev.com/174857165e6d4ba57d36c8c5ce10134ebf8fcd9a
Cr-Commit-Position: refs/heads/master@{#367465}

[Xianzhu, 2015-12-29 20:17:26]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

[Xianzhu, 2015-12-29 20:17:26]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[Xianzhu, 2015-12-29 20:17:31]

wangxianzhu@chromium.org changed reviewers:
+ chrishtr@chromium.org, schenney@chromium.org

[Xianzhu, 2015-12-29 20:18:54]

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

[commit-bot: I haz the power, 2015-12-29 20:19:08]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1552693002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1552693002/1

[commit-bot: I haz the power, 2015-12-29 21:52:05]

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

[commit-bot: I haz the power, 2015-12-29 21:52:07]

Dry run: This issue passed the CQ dry run.

[chrishtr, 2016-01-04 16:55:58]

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/public/p...
File third_party/WebKit/public/platform/WebContentLayerClient.h (right):

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/public/p...
third_party/WebKit/public/platform/WebContentLayerClient.h:43:
PaintDefaultBehavior,
Is there any point to PaintDefaultBehavior in sync paint mode? If not, how about
we change it to PaintAllContent and have it behave like
SubsequenceCachingDisabled?

[Xianzhu, 2016-01-04 17:19:17]

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/public/p...
File third_party/WebKit/public/platform/WebContentLayerClient.h (right):

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/public/p...
third_party/WebKit/public/platform/WebContentLayerClient.h:43:
PaintDefaultBehavior,
On 2016/01/04 16:55:58, chrishtr wrote:
> Is there any point to PaintDefaultBehavior in sync paint mode? If not, how
about
> we change it to PaintAllContent and have it behave like
> SubsequenceCachingDisabled?

We still need PaintDefaultBehavior for non-benchmarking mode. For benchmarking,
it seems to be useful to measure the cost other than painting, e.g. interest
rect calculation and paintArtifactToWebDisplayItemList().

[chrishtr, 2016-01-04 17:32:08]

lgtm

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/graphics/paint/PaintController.h
(right):

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/graphics/paint/PaintController.h:133: bool
subsequenceCachingIsDisabled() const { return m_subsequenceCachingDisabled; }
Add a comment that these are used only benchmarking. Group with lines 130-131 to
indicate that
for those lines too.

[Xianzhu, 2016-01-04 17:39:55]

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/Source/p...
File third_party/WebKit/Source/platform/graphics/paint/PaintController.h
(right):

https://codereview.chromium.org/1552693002/diff/1/third_party/WebKit/Source/p...
third_party/WebKit/Source/platform/graphics/paint/PaintController.h:133: bool
subsequenceCachingIsDisabled() const { return m_subsequenceCachingDisabled; }
On 2016/01/04 17:32:08, chrishtr wrote:
> Add a comment that these are used only benchmarking. Group with lines 130-131
to
> indicate that
> for those lines too.

Done.

[Xianzhu, 2016-01-04 17:40:13]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

[Xianzhu, 2016-01-04 19:27:26]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

[Xianzhu, 2016-01-04 19:27:57]

wangxianzhu@chromium.org changed reviewers:
+ nednguyen@google.com

[Xianzhu, 2016-01-04 19:27:58]

+nednguyen@ for tools/perf

[Xianzhu, 2016-01-04 19:32:54]

wangxianzhu@chromium.org changed reviewers:
+ ajuma@chromium.org

[Xianzhu, 2016-01-04 19:32:54]

+ajuma@ for cc/

[ajuma, 2016-01-04 19:41:07]

On 2016/01/04 19:32:54, Xianzhu wrote:
> +ajuma@ for cc/

cc lgtm.

[nednguyen, 2016-01-04 20:24:52]

On 2016/01/04 19:41:07, ajuma wrote:
> On 2016/01/04 19:32:54, Xianzhu wrote:
> > +ajuma@ for cc/
> 
> cc lgtm.

How stable are the additional metrics?

[nednguyen, 2016-01-04 20:24:54]

On 2016/01/04 19:41:07, ajuma wrote:
> On 2016/01/04 19:32:54, Xianzhu wrote:
> > +ajuma@ for cc/
> 
> cc lgtm.

How stable are the additional metrics?

[Xianzhu, 2016-01-04 20:37:29]

On 2016/01/04 20:24:54, nednguyen (ooo till 1-3) wrote:
> On 2016/01/04 19:41:07, ajuma wrote:
> > On 2016/01/04 19:32:54, Xianzhu wrote:
> > > +ajuma@ for cc/
> > 
> > cc lgtm.
> 
> How stable are the additional metrics?

Record_time metrics are quite stable because we choose the highest performance
from 100 runs. The new metric is the same as record_time before synchronized
painting was enabled. Sample graph is
https://chromeperf.appspot.com/report?sid=e38127ee99af815f1f071dcf31fefeb8b68....
The up and down occurred when I enabled/disabled the synchronized painting flag.

[nednguyen, 2016-01-04 21:02:32]

lgtm

[Xianzhu, 2016-01-04 21:20:58]

The CQ bit was checked by wangxianzhu@chromium.org

[Xianzhu, 2016-01-04 21:20:59]

The patchset sent to the CQ was uploaded after l-g-t-m from
chrishtr@chromium.org
Link to the patchset: https://codereview.chromium.org/1552693002/#ps40001
(title: "Rebase on origin/master")

[commit-bot: I haz the power, 2016-01-04 21:21:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1552693002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1552693002/40001

[commit-bot: I haz the power, 2016-01-04 21:35:04]

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

[commit-bot: I haz the power, 2016-01-04 21:35:05]

Try jobs failed on following builders:
  linux_perf_bisect on tryserver.chromium.perf (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.perf/builders/linux_perf_bisec...)

[Xianzhu, 2016-01-04 21:42:17]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-01-04 21:43:21]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1552693002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1552693002/40001

[commit-bot: I haz the power, 2016-01-04 21:59:55]

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

[commit-bot: I haz the power, 2016-01-04 21:59:56]

Try jobs failed on following builders:
  linux_perf_bisect on tryserver.chromium.perf (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.perf/builders/linux_perf_bisec...)

[Xianzhu, 2016-01-04 22:45:31]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-01-04 22:46:51]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1552693002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1552693002/40001

[commit-bot: I haz the power, 2016-01-04 23:00:30]

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

[commit-bot: I haz the power, 2016-01-04 23:00:31]

Try jobs failed on following builders:
  linux_perf_bisect on tryserver.chromium.perf (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.perf/builders/linux_perf_bisec...)

[Xianzhu, 2016-01-05 01:44:03]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-01-05 01:44:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1552693002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1552693002/40001

[commit-bot: I haz the power, 2016-01-05 02:48:43]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

[commit-bot: I haz the power, 2016-01-05 02:48:44]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-01-05 02:49:59]

Description was changed from

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect
==========

to

==========
Add paint testing mode subsequence_caching_disabled

With synchronized painting, because painting is already done when
testing record_time, the normal record_time result doesn't contain
anything for painting.

Add subsequence_caching_disabled to force painting with subsequence
caching disabled, to measure the performance of display item caching.
The new mode should generates the same record_time before
synchronized painting is enabled.

The default record_time is still useful to measure the cost other
than paint.

BUG=536999
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
CQ_EXTRA_TRYBOTS=tryserver.chromium.perf:linux_perf_bisect;tryserver.chromium.perf:mac_10_10_perf_bisect;tryserver.chromium.perf:win_perf_bisect

Committed: https://crrev.com/174857165e6d4ba57d36c8c5ce10134ebf8fcd9a
Cr-Commit-Position: refs/heads/master@{#367465}
==========

[commit-bot: I haz the power, 2016-01-05 02:50:00]

Patchset 3 (id:??) landed as
https://crrev.com/174857165e6d4ba57d36c8c5ce10134ebf8fcd9a
Cr-Commit-Position: refs/heads/master@{#367465}