cc: Optimize TileManager::CleanUpReleasedTiles.

cc: Optimize TileManager::CleanUpReleasedTiles.

This patch ensures that we don't do extra hidden iterations in the
loop over tiles (erase is linear).

R=danakj, enne

Committed: https://crrev.com/d673ee5f972c3ce695d8a9fb10ac7a03ccc01578
Cr-Commit-Position: refs/heads/master@{#324161}

[vmpstr, 2015-04-07 22:58:37]

Please take a look.

[danakj, 2015-04-07 23:01:08]

LGTM

[vmpstr, 2015-04-07 23:04:34]

The CQ bit was checked by vmpstr@chromium.org

[enne (OOO), 2015-04-07 23:04:48]

https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.cc
File cc/resources/tile_manager.cc (right):

https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.c...
cc/resources/tile_manager.cc:280: std::vector<Tile*> tiles_to_retain;
Should you reserve this to be the number of tiles in released_tiles to avoid
unneeded allocations?

[commit-bot: I haz the power, 2015-04-07 23:05:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1069493003/1

[vmpstr, 2015-04-07 23:10:24]

https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.cc
File cc/resources/tile_manager.cc (right):

https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.c...
cc/resources/tile_manager.cc:280: std::vector<Tile*> tiles_to_retain;
On 2015/04/07 23:04:48, enne wrote:
> Should you reserve this to be the number of tiles in released_tiles to avoid
> unneeded allocations?

I'm not sure it's worth it, since the number of tiles that have a raster task is
bounded by a small number typically, and push_back is amortized constant. Number
of released tiles is technically not bounded on the other hand.

[enne (OOO), 2015-04-07 23:13:52]

On 2015/04/07 at 23:10:24, vmpstr wrote:
> https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.cc
> File cc/resources/tile_manager.cc (right):
> 
>
https://codereview.chromium.org/1069493003/diff/1/cc/resources/tile_manager.c...
> cc/resources/tile_manager.cc:280: std::vector<Tile*> tiles_to_retain;
> On 2015/04/07 23:04:48, enne wrote:
> > Should you reserve this to be the number of tiles in released_tiles to avoid
> > unneeded allocations?
> 
> I'm not sure it's worth it, since the number of tiles that have a raster task
is bounded by a small number typically, and push_back is amortized constant.
Number of released tiles is technically not bounded on the other hand.

Gotcha.  If in practice this is also like <= 32, then whatever.  lgtm

[commit-bot: I haz the power, 2015-04-08 00:27:14]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-04-08 00:28:01]

Patchset 1 (id:??) landed as
https://crrev.com/d673ee5f972c3ce695d8a9fb10ac7a03ccc01578
Cr-Commit-Position: refs/heads/master@{#324161}