Turn on NVPR 4x MSAA by default when supported in DM and nanobench.

Turn on NVPR 4x MSAA by default when supported in DM and nanobench.

This brings DM and nanobench's default configs in line with GM's.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/31f88675718966bbb7f09718b40de10c7e214739

[mtklein, 2014-11-05 16:44:05]

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

[bsalomon, 2014-11-05 19:52:31]

lgtm

[mtklein, 2014-11-05 20:19:03]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-11-05 20:20:21]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/704563003/20001

[commit-bot: I haz the power, 2014-11-05 20:28:29]

Committed patchset #2 (id:20001) as 31f88675718966bbb7f09718b40de10c7e214739

[egdaniel, 2014-11-05 22:20:17]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/693933004/ by egdaniel@google.com.

The reason for reverting is: Timing out on nvpr when drawing conics. Fix will
take some thought so reverting for now..

[Kimmo Kinnunen, 2014-11-11 07:14:01]

On 2014/11/05 22:20:17, egdaniel wrote:
> A revert of this CL (patchset #2 id:20001) has been created in
> https://codereview.chromium.org/693933004/ by mailto:egdaniel@google.com.
> 
> The reason for reverting is: Timing out on nvpr when drawing conics. Fix will
> take some thought so reverting for now..

Hey, I could try to fix the underlying problem. I wonder which test caused the
timeout?
I assume https://codereview.chromium.org/709473002 fixed the problem? If so, do
I understand correctly that the test itself may not be slow, but nanobench fails
to converge, thus just looping very long time (which caused the original
overflow)?

[Kimmo Kinnunen, 2014-11-11 12:51:49]

On 2014/11/11 07:14:01, Kimmo Kinnunen wrote:
> Hey, I could try to fix the underlying problem. I wonder which test caused the
> timeout?

Nevermind, if it was path_hairline_{small,big}_AA_conic, I found it.

[bsalomon, 2014-11-11 12:54:47]

On 2014/11/11 12:51:49, Kimmo Kinnunen wrote:
> On 2014/11/11 07:14:01, Kimmo Kinnunen wrote:
> > Hey, I could try to fix the underlying problem. I wonder which test caused
the
> > timeout?
> 
> Nevermind, if it was path_hairline_{small,big}_AA_conic, I found it.

I'm pretty sure it was those tests.

[mtklein, 2014-11-11 13:11:59]

On 2014/11/11 12:54:47, bsalomon wrote:
> On 2014/11/11 12:51:49, Kimmo Kinnunen wrote:
> > On 2014/11/11 07:14:01, Kimmo Kinnunen wrote:
> > > Hey, I could try to fix the underlying problem. I wonder which test caused
> the
> > > timeout?
> > 
> > Nevermind, if it was path_hairline_{small,big}_AA_conic, I found it.
> 
> I'm pretty sure it was those tests.

Right.  IIRC, the bench was taking roughly a constant time to run (drawing
nothing) no matter how nanobench tuned the loops.