[Telemetry] Improve unittest cycle time and fix testIFrame.

[Telemetry] Improve unittest cycle time and fix testIFrame.

Starting chrome to about:blank instead of the NTP improves telemetry_unittests
cycle time by ~14% (172s -> 148s) on my MBP.

This required a couple of test updates:
1. It made testIFrame fail deterministically which exposed the source of
flake. The context_id isn't the current index, it is just a monotonically
increasing ID. So it depends on the number of contexts used previously.
2. It made the GPU memory test measure a larger delta (presumably due to the simpler previous page).

BUG=394454, 388256
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=289423

[dtu, 2014-08-13 03:49:55]

The CQ bit was checked by dtu@chromium.org

[dtu, 2014-08-13 03:49:55]

lgtm

[commit-bot: I haz the power, 2014-08-13 03:51:40]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/tonyg@chromium.org/466233002/1

[achuithb, 2014-08-13 07:57:34]

LGTM. Thanks for fixing this!

[commit-bot: I haz the power, 2014-08-13 08:02:26]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  mac_gpu_retina_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/mac_gpu_retina_tr...)

[commit-bot: I haz the power, 2014-08-13 08:23:03]

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

[commit-bot: I haz the power, 2014-08-13 08:23:04]

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

[tonyg, 2014-08-13 13:34:58]

The CQ bit was checked by tonyg@chromium.org

[commit-bot: I haz the power, 2014-08-13 13:36:18]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/tonyg@chromium.org/466233002/1

[commit-bot: I haz the power, 2014-08-13 13:56:51]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  mac_gpu_retina_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/mac_gpu_retina_tr...)

[tonyg, 2014-08-13 14:10:18]

Hi zmo, kbr,

This patch made the retina gpu memory test fail with:
Failure: Memory allocation too high (was 204 MB, should be 192 MB +/- 8 MB)

My suspicion is that it failed because about:blank is so much simpler than the
NTP, and thus the test measures a larger delta on the first real page. Starting
to blank should be more stable and faster in the long run, so I think this is
the right thing to do. So, I've updated the patch to increase the wiggle room.
Alternatively I could modify the benchmark to continue to start to the NTP if
you prefer.

Please take a look.

-Tony

http://build.chromium.org/p/tryserver.chromium.gpu/builders/mac_gpu_retina_tr...

[Zhenyao Mo, 2014-08-13 19:42:55]

LGTM

I am wondering instead of changing the WIGGLE_ROOM, maybe we should adjust the
base (MEMORY_LIMIT) with this change?

I think landing as is should be OK.

[tonyg, 2014-08-13 19:43:34]

The CQ bit was checked by tonyg@chromium.org

[commit-bot: I haz the power, 2014-08-13 19:45:40]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/tonyg@chromium.org/466233002/20001

[commit-bot: I haz the power, 2014-08-14 00:27:58]

Committed patchset #2 (20001) as 289423