Add experimental v8 code cache options.

Add experimental v8 code cache options.

- heuristics-default: bail out to default behavior.
- heuristics-default-mobile: as above, tuned for mobile.
- recent: only cache recently compiled scripts.
- recent-small: only cache small recently compiled scipts.

R=jochen@chromium.org, vogelheim@chromium.org
BUG=chromium:399580
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=188675

[vogelheim, 2015-01-20 12:25:21]

lgtm

Thanks, looks good!

[jochen (gone - plz use gerrit), 2015-01-20 13:08:22]

lgtm

[Yang, 2015-01-20 13:09:35]

The CQ bit was checked by yangguo@chromium.org

[commit-bot: I haz the power, 2015-01-20 13:10:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/845923003/1

[marja, 2015-01-20 13:14:32]

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

[marja, 2015-01-20 13:14:33]

Did I understand this correctly?

https://codereview.chromium.org/845923003/diff/1/Source/bindings/core/v8/V8Sc...
File Source/bindings/core/v8/V8ScriptRunner.cpp (right):

https://codereview.chromium.org/845923003/diff/1/Source/bindings/core/v8/V8Sc...
Source/bindings/core/v8/V8ScriptRunner.cpp:313: if
(!isResourceHotForCaching(resource)) {
So this caches the code if we have recently compiled it, right? So it first
compiles w/o caching, then compiles again, and if the first compilation was
recent enough, it now puts data to the cache.

[Yang, 2015-01-20 13:33:10]

On 2015/01/20 13:14:33, marja wrote:
> Did I understand this correctly?
> 
>
https://codereview.chromium.org/845923003/diff/1/Source/bindings/core/v8/V8Sc...
> File Source/bindings/core/v8/V8ScriptRunner.cpp (right):
> 
>
https://codereview.chromium.org/845923003/diff/1/Source/bindings/core/v8/V8Sc...
> Source/bindings/core/v8/V8ScriptRunner.cpp:313: if
> (!isResourceHotForCaching(resource)) {
> So this caches the code if we have recently compiled it, right? So it first
> compiles w/o caching, then compiles again, and if the first compilation was
> recent enough, it now puts data to the cache.

Exactly. The first time we end up with no caching at all, but set the time
stamp. The second time we compare the time stamps to decide whether we want to
cache. If yes, we serialize/cache. If not, we reset the time stamp for next
time.

[marja, 2015-01-20 13:40:05]

I'm worried that this wouldn't be hit often enough to actually put the data into
the cache often enough to show an effect... as a follow up you might want to add
UMA to track how often the stuff is actually put into the cache / taken from the
cache.

[Yang, 2015-01-20 13:41:00]

On 2015/01/20 13:40:05, marja wrote:
> I'm worried that this wouldn't be hit often enough to actually put the data
into
> the cache often enough to show an effect... as a follow up you might want to
add
> UMA to track how often the stuff is actually put into the cache / taken from
the
> cache.

The V8.CompileSerialize and V8.CompileDeserialize histograms can also act as
counters :)

[marja, 2015-01-20 13:42:14]

Ah right, then you don't need anything extra for this, nevermind :)

[commit-bot: I haz the power, 2015-01-20 14:57:10]

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

[commit-bot: I haz the power, 2015-01-20 14:57:11]

Try jobs failed on following builders:
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/46948)

[Yang, 2015-01-20 14:58:36]

The CQ bit was checked by yangguo@chromium.org

[commit-bot: I haz the power, 2015-01-20 14:59:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/845923003/1

[commit-bot: I haz the power, 2015-01-20 15:47:12]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=188675