nanobench: split CPU and GPU .SKP tile sizes.

nanobench: split CPU and GPU .SKP tile sizes.

Let's make CPU-bound .SKP benching mimic Chrome's tiles.
Unfortunately, the CPU code also performs a lot better with those big wide tiles...

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/bf9e60006987f23773b1e0bdfad9d424d4bc5170

[mtklein_C, 2015-06-16 16:54:38]

mtklein@chromium.org changed reviewers:
+ robertphillips@google.com

[mtklein_C, 2015-06-16 16:56:30]

+CC herb,

I think this is why we're not seeing glyph cache contention to be a problem when
benching nytimes.  We were only drawing 2 big tiles.
This CL breaks it up into 16 256x256 tiles, which is more like how Chrome would
do it.

[mtklein_C, 2015-06-16 16:56:38]

The CQ bit was checked by mtklein@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-06-16 16:56:51]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1189863002/1

[mtklein, 2015-06-16 16:58:27]

mtklein@google.com changed reviewers:
+ mtklein@google.com

[mtklein, 2015-06-16 16:58:27]

(I say more like because the real logic is somewhat more interesting:
https://code.google.com/p/chromium/codesearch#chromium/src/content/renderer/g...
)

[commit-bot: I haz the power, 2015-06-16 17:00:30]

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

[commit-bot: I haz the power, 2015-06-16 17:00:31]

Dry run: Try jobs failed on following builders:
  Build-Win-MSVC-x86-Debug-Trybot on client.skia.compile (JOB_FAILED,
http://build.chromium.org/p/client.skia.compile/builders/Build-Win-MSVC-x86-D...)
  Build-Win-MSVC-x86_64-Debug-Trybot on client.skia.compile (JOB_FAILED,
http://build.chromium.org/p/client.skia.compile/builders/Build-Win-MSVC-x86_6...)

[mtklein, 2015-06-16 17:03:14]

The CQ bit was checked by mtklein@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-06-16 17:03:30]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1189863002/20001

[commit-bot: I haz the power, 2015-06-16 17:08:40]

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

[commit-bot: I haz the power, 2015-06-16 17:08:41]

Dry run: This issue passed the CQ dry run.

[robertphillips, 2015-06-16 17:26:17]

lgtm. Do we use the old command line args anywhere?

[mtklein, 2015-06-16 17:35:44]

Ooh, good catch.  Yes, we did set back to 256x256 on HD2000.  Updated.

[mtklein, 2015-06-16 17:36:07]

The CQ bit was checked by mtklein@google.com

[mtklein, 2015-06-16 17:36:07]

The patchset sent to the CQ was uploaded after l-g-t-m from
robertphillips@google.com
Link to the patchset: https://codereview.chromium.org/1189863002/#ps40001
(title: "update flags")

[commit-bot: I haz the power, 2015-06-16 17:36:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1189863002/40001

[commit-bot: I haz the power, 2015-06-16 17:41:32]

Committed patchset #3 (id:40001) as
https://skia.googlesource.com/skia/+/bf9e60006987f23773b1e0bdfad9d424d4bc5170

[mtklein, 2015-06-16 20:32:10]

+Eric

Hey Sheriff, 

FYI, this CL is expected to totally wreck perf.skia.org.  It'll affect all the
SKPs on all the CPU bots.  Some bots will go faster, some bots will go slower. 
Some SKPs will go faster, some SKPs will go slower.

Interestingly, the arithmetic mean runtime has gone up, but the geometric mean
runtime has gone down: https://perf.skia.org/#3633.  I think that means it's
generally faster but has introduced a few severe performance regressions.

If perf alerts come up I'll try to triage them, but it should be fine to just
tag them all ignored "mtklein says so" if I miss any.