Update a result vector when sampling animation effects.

Update a result vector when sampling animation effects.

Switch AnimationEffect::sample() from being functional and return a
new result vector to instead take a result vector to update. This lets animation
loops that repeatedly sample to reuse their vector allocations.

R=dstockwell,shans
BUG=
NOTRY=true

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185840

[sof, 2014-11-19 17:01:34]

sigbjornf@opera.com changed reviewers:
+ oilpan-reviews@chromium.org

[sof, 2014-11-19 17:01:34]

Please take a look.

This reduces the allocation rates for some of the blink_perf.animation tests,
which I think is in our general interest.

With Oilpan enabled, seeing a 6% improvement on balls-keyframe-animations.html
compared to ToT (linux).

Adding oilpan-reviews as initial reviewers, to just check that this is an
optimization to pursue, i.e., avoid excessive allocations. wdyt?

[sof, 2014-11-19 22:33:18]

blink_perf.animation results:
https://docs.google.com/spreadsheets/d/1OzLWVacJZ1Qj59ttmJzM815mlvRsd7Wv79qMQ...

As regards allocation counts, for balls-keyframe-animation.html,
InterpolationEffect::getActiveInterpolations() will with ToT allocate 525,000
vectors. With this CL, 5000 vectors are allocated by the same method.

[haraken, 2014-11-20 01:33:07]

On 2014/11/19 22:33:18, sof wrote:
> blink_perf.animation results:
>
https://docs.google.com/spreadsheets/d/1OzLWVacJZ1Qj59ttmJzM815mlvRsd7Wv79qMQ...
> 
> As regards allocation counts, for balls-keyframe-animation.html,
> InterpolationEffect::getActiveInterpolations() will with ToT allocate 525,000
> vectors. With this CL, 5000 vectors are allocated by the same method.

Looks great!

[sof, 2014-11-20 06:11:15]

sigbjornf@opera.com changed reviewers:
+ ericwilligers@chromium.org, shans@chromium.org

[sof, 2014-11-20 06:11:16]

+shans, ericwilligers as reviewers.

[dstockwell, 2014-11-20 09:06:08]

dstockwell@chromium.org changed reviewers:
+ dstockwell@chromium.org
- ericwilligers@chromium.org

[dstockwell, 2014-11-20 09:06:09]

lgtm, but wait for shans@ to take a look

[sof, 2014-11-21 15:06:28]

On 2014/11/20 09:06:09, dstockwell wrote:
> lgtm, but wait for shans@ to take a look

certainly.. :)

[shans, 2014-11-22 11:06:47]

On 2014/11/21 15:06:28, sof wrote:
> On 2014/11/20 09:06:09, dstockwell wrote:
> > lgtm, but wait for shans@ to take a look
> 
> certainly.. :)

lgtm too, sorry for taking this long to get back to you.

[sof, 2014-11-22 15:17:21]

On 2014/11/22 11:06:47, shans wrote:
> On 2014/11/21 15:06:28, sof wrote:
> > On 2014/11/20 09:06:09, dstockwell wrote:
> > > lgtm, but wait for shans@ to take a look
> > 
> > certainly.. :)
> 
> lgtm too, sorry for taking this long to get back to you.

nw, thanks both for the reviews. Let's see if this brings measurable perf
improvements..

[sof, 2014-11-22 15:17:37]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-11-22 15:17:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/732153003/1

[commit-bot: I haz the power, 2014-11-22 15:56:15]

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

[commit-bot: I haz the power, 2014-11-22 15:56:16]

Try jobs failed on following builders:
  linux_gpu_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/linux_gpu_trigger...)

[sof, 2014-11-22 16:17:59]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-11-22 16:18:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/732153003/1

[commit-bot: I haz the power, 2014-11-22 16:19:14]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=185840