Remove blink::RuntimeEnabledFeatures::slimmingPaintDisplayItemCache

Remove blink::RuntimeEnabledFeatures::slimmingPaintDisplayItemCache

The flag was added to allow us test the mode with display item cache
enabled or disabled. It seems that we haven't been using it for some
time, so perhaps it's time to remove it and always enable the cache.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=192218

[Xianzhu, 2015-03-18 17:37:33]

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

[Stephen Chennney, 2015-03-18 17:43:01]

I think it's time to remove it, and the code changes look fine. I'll let pdr
give the final approval in case the rest of the team thinks otherwise.

[chrishtr, 2015-03-18 17:45:53]

chrishtr@chromium.org changed reviewers:
+ chrishtr@chromium.org

[chrishtr, 2015-03-18 17:45:54]

I observed yesterday that resizing the window results in broken output. Turning
off the cache appears to fix this. Is this bug tracked yet? I'd like it to be
tracked and fixed before removing this flag.

[pdr., 2015-03-18 18:06:47]

On 2015/03/18 at 17:45:54, chrishtr wrote:
> I observed yesterday that resizing the window results in broken output.
Turning off the cache appears to fix this. Is this bug tracked yet? I'd like it
to be tracked and fixed before removing this flag.

Would it be interesting to kick off a run of the top 10k webpages with this flag
disabled before we remove it?

[Xianzhu, 2015-03-18 19:05:40]

On 2015/03/18 18:06:47, pdr wrote:
> On 2015/03/18 at 17:45:54, chrishtr wrote:
> > I observed yesterday that resizing the window results in broken output.
> Turning off the cache appears to fix this. Is this bug tracked yet? I'd like
it
> to be tracked and fixed before removing this flag.
> 
> Would it be interesting to kick off a run of the top 10k webpages with this
flag
> disabled before we remove it?

I'd like to see the results. The change of no_cache and no_paint results can
show the cost of the merge algorithm.

[chrishtr, 2015-03-18 19:17:50]

On 2015/03/18 at 19:05:40, wangxianzhu wrote:
> On 2015/03/18 18:06:47, pdr wrote:
> > On 2015/03/18 at 17:45:54, chrishtr wrote:
> > > I observed yesterday that resizing the window results in broken output.
> > Turning off the cache appears to fix this. Is this bug tracked yet? I'd like
it
> > to be tracked and fixed before removing this flag.
> > 
> > Would it be interesting to kick off a run of the top 10k webpages with this
flag
> > disabled before we remove it?
> 
> I'd like to see the results. The change of no_cache and no_paint results can
show the cost of the merge algorithm.

+1. It's easy to do a run on cluster telemetry, let me know if you need a
pointer.

[Xianzhu, 2015-03-18 19:19:32]

On 2015/03/18 19:17:50, chrishtr wrote:
> On 2015/03/18 at 19:05:40, wangxianzhu wrote:
> > On 2015/03/18 18:06:47, pdr wrote:
> > > On 2015/03/18 at 17:45:54, chrishtr wrote:
> > > > I observed yesterday that resizing the window results in broken output.
> > > Turning off the cache appears to fix this. Is this bug tracked yet? I'd
like
> it
> > > to be tracked and fixed before removing this flag.
> > > 
> > > Would it be interesting to kick off a run of the top 10k webpages with
this
> flag
> > > disabled before we remove it?
> > 
> > I'd like to see the results. The change of no_cache and no_paint results can
> show the cost of the merge algorithm.
> 
> +1. It's easy to do a run on cluster telemetry, let me know if you need a
> pointer.

Yes, I need a pointer :)

[pdr., 2015-03-18 19:21:21]

On 2015/03/18 at 19:19:32, wangxianzhu wrote:
> On 2015/03/18 19:17:50, chrishtr wrote:
> > On 2015/03/18 at 19:05:40, wangxianzhu wrote:
> > > On 2015/03/18 18:06:47, pdr wrote:
> > > > On 2015/03/18 at 17:45:54, chrishtr wrote:
> > > > > I observed yesterday that resizing the window results in broken
output.
> > > > Turning off the cache appears to fix this. Is this bug tracked yet? I'd
like
> > it
> > > > to be tracked and fixed before removing this flag.
> > > > 
> > > > Would it be interesting to kick off a run of the top 10k webpages with
this
> > flag
> > > > disabled before we remove it?
> > > 
> > > I'd like to see the results. The change of no_cache and no_paint results
can
> > show the cost of the merge algorithm.
> > 
> > +1. It's easy to do a run on cluster telemetry, let me know if you need a
> > pointer.
> 
> Yes, I need a pointer :)

I'll run this for you. Results coming up in ~4hrs

[Xianzhu, 2015-03-18 19:25:42]

On 2015/03/18 19:21:21, pdr wrote:
> On 2015/03/18 at 19:19:32, wangxianzhu wrote:
> > On 2015/03/18 19:17:50, chrishtr wrote:
> > > On 2015/03/18 at 19:05:40, wangxianzhu wrote:
> > > > On 2015/03/18 18:06:47, pdr wrote:
> > > > > On 2015/03/18 at 17:45:54, chrishtr wrote:
> > > > > > I observed yesterday that resizing the window results in broken
> output.
> > > > > Turning off the cache appears to fix this. Is this bug tracked yet?
I'd
> like
> > > it
> > > > > to be tracked and fixed before removing this flag.
> > > > > 
> > > > > Would it be interesting to kick off a run of the top 10k webpages with
> this
> > > flag
> > > > > disabled before we remove it?
> > > > 
> > > > I'd like to see the results. The change of no_cache and no_paint results
> can
> > > show the cost of the merge algorithm.
> > > 
> > > +1. It's easy to do a run on cluster telemetry, let me know if you need a
> > > pointer.
> > 
> > Yes, I need a pointer :)
> 
> I'll run this for you. Results coming up in ~4hrs

Thanks!

[pdr., 2015-03-18 19:26:48]

On 2015/03/18 at 19:21:21, pdr wrote:
> On 2015/03/18 at 19:19:32, wangxianzhu wrote:
> > On 2015/03/18 19:17:50, chrishtr wrote:
> > > On 2015/03/18 at 19:05:40, wangxianzhu wrote:
> > > > On 2015/03/18 18:06:47, pdr wrote:
> > > > > On 2015/03/18 at 17:45:54, chrishtr wrote:
> > > > > > I observed yesterday that resizing the window results in broken
output.
> > > > > Turning off the cache appears to fix this. Is this bug tracked yet?
I'd like
> > > it
> > > > > to be tracked and fixed before removing this flag.
> > > > > 
> > > > > Would it be interesting to kick off a run of the top 10k webpages with
this
> > > flag
> > > > > disabled before we remove it?
> > > > 
> > > > I'd like to see the results. The change of no_cache and no_paint results
can
> > > show the cost of the merge algorithm.
> > > 
> > > +1. It's easy to do a run on cluster telemetry, let me know if you need a
> > > pointer.
> > 
> > Yes, I need a pointer :)
> 
> I'll run this for you. Results coming up in ~4hrs

You're now in the queue.

What I do is create a patch disabling the feature
(https://codereview.chromium.org/download/issue1015203002_1.diff) and then paste
it in the blink section of
http://skia-tree-status.appspot.com/skia-telemetry/chromium_try. To make sure
slimming paint is enabled, I click "show details" and add
--enable-slimming-paint to the with and without flags. We should have results in
a little over 5 hours because this morning's slimming paint task is still
running.

[Stephen Chennney, 2015-03-18 19:37:52]

You can also apply my "always turn on slimming-paint" patch in the Chromium
section. https://codereview.chromium.org/1003273003/

[pdr., 2015-03-19 00:10:08]

The data is in, and it's interesting!
https://docs.google.com/document/d/1moskT1Tkg8GcHZylRNQakyrCjk3LR_O5Yys9fDriV...

Lets discuss on the doc.

[chrishtr, 2015-03-19 05:14:31]

This performance data is very interesting. I'm uncomfortably excited.

[Xianzhu, 2015-03-19 17:59:27]

Thanks pdr@ for running the benchmarks. It's really useful and we discovered
some interesting facts.

However, we also discovered that the visible_content_rect passed from the
benchmark is not suitable for the cache-disabled mode which requires full-layer
recording. The data collected with partial-layer recording in cache-disabled
mode is inaccurate (though, we did find some interesting facts).

I think we are ready to remove the flag.

This CL requires chromium-side change
https://codereview.chromium.org/1012383005/.

[chrishtr, 2015-03-19 18:40:24]

Is the issue with window resize solved completely?

[Xianzhu, 2015-03-19 19:00:43]

On 2015/03/19 18:40:24, chrishtr wrote:
> Is the issue with window resize solved completely?

It will be resolved by https://codereview.chromium.org/1011393003/.

[chrishtr, 2015-03-19 19:44:13]

lgtm

[Xianzhu, 2015-03-19 22:14:30]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2015-03-19 22:14:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1018053003/20001

[commit-bot: I haz the power, 2015-03-19 22:17:44]

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

[commit-bot: I haz the power, 2015-03-19 22:17:45]

Try jobs failed on following builders:
  mac_blink_compile_dbg on tryserver.blink (JOB_FAILED,
http://build.chromium.org/p/tryserver.blink/builders/mac_blink_compile_dbg/bu...)

[Xianzhu, 2015-03-19 23:28:41]

The CQ bit was checked by wangxianzhu@chromium.org

[Xianzhu, 2015-03-19 23:28:42]

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/1018053003/#ps40001
(title: "Rebase")

[commit-bot: I haz the power, 2015-03-19 23:29:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1018053003/40001

[commit-bot: I haz the power, 2015-03-20 01:49:22]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=192218