Run more tests on the Webkit Layout Valgrind bot and do it in parallel

Run more tests on the Webkit Layout Valgrind bot and do it in parallel
Some numbers: with --experimental-fully-parallel my Z600 (16 threads)
runs 360 tests in 30 minutes. The bot has 4 vcores, we may ask for more if it
becomes too slow.
I haven't seen any test failures locally; probably thanks to --run-singly.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109787

[Lei Zhang, 2011-11-11 22:29:08]

LGTM

ojan: how "experimental" is --experimental-fully-parallel ? Should we be using
it?

[ojan, 2011-11-11 22:44:34]

On 2011/11/11 22:29:08, Lei Zhang wrote:
> LGTM
> 
> ojan: how "experimental" is --experimental-fully-parallel ? Should we be using
> it?

Last I checked, it increases flakiness a lot. The valgrind tests have so many
other causes of flakiness that I doubt this will make things much worse and is
probably worth the benefit.

[commit-bot: I haz the power, 2011-11-12 06:57:07]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/timurrrr@chromium.org/8463012/1

[commit-bot: I haz the power, 2011-11-12 08:10:44]

Change committed as 109787

[Timur Iskhodzhanov, 2011-11-12 19:55:03]

FTR: As expected, the cycle time is now ~1h for 200 tests.
before the change, we were running 60 tests in ~40m on average.

btw, most of the tests pass :)