Adding new bench tests for lighting and blur

Adding new bench tests for lighting and blur

Committed: https://code.google.com/p/skia/source/detail?r=8846

[Stephen White, 2013-04-24 17:35:56]

https://codereview.chromium.org/14474004/diff/1/bench/BlurBench.cpp
File bench/BlurBench.cpp (right):

https://codereview.chromium.org/14474004/diff/1/bench/BlurBench.cpp#newcode30
bench/BlurBench.cpp:30: class BlurMaskBench : public SkBenchmark {
I think we should probably keep this as BlurBench, since the names in the graphs
will still be blur_foo.  Maybe the new one could be BlurImageFilterBench. 
Unless there's a lot of opportunity for code sharing in the benches, I'd suggest
we put the new one in a new file (say, BlurImageFilterBench.cpp).

https://codereview.chromium.org/14474004/diff/1/bench/BlurBench.cpp#newcode86
bench/BlurBench.cpp:86: BlurBench(void* param, float sigmaX, float sigmaY,  bool
small) :
For symmetry, we should probably make these parameters SkScalars, and #define
the filter widths with SkFloatToScalar, as BlurBench does.

https://codereview.chromium.org/14474004/diff/1/bench/BlurBench.cpp#newcode91
bench/BlurBench.cpp:91: fName.printf("blur_small_%.2f_%.2f", sigmaX, sigmaY);
If we rename the class to BlurImageFilterBench, this should be
blur_image_filter_small, so it's clear in the graphing output which filter is
being tested.

https://codereview.chromium.org/14474004/diff/1/bench/LightingBench.cpp
File bench/LightingBench.cpp (right):

https://codereview.chromium.org/14474004/diff/1/bench/LightingBench.cpp#newco...
bench/LightingBench.cpp:166: SkScalar azimuthRad =
SkDegreesToRadians(SkIntToScalar(225));
Nit:  looks like there are lots of opportunities for code sharing here with the
Diffuse benches.  You can do that in a followup patch if you like.

[sugoi1, 2013-04-24 18:49:10]

https://codereview.chromium.org/14474004/diff/1/bench/LightingBench.cpp
File bench/LightingBench.cpp (right):

https://codereview.chromium.org/14474004/diff/1/bench/LightingBench.cpp#newco...
bench/LightingBench.cpp:166: SkScalar azimuthRad =
SkDegreesToRadians(SkIntToScalar(225));
On 2013/04/24 17:35:56, Stephen White wrote:
> Nit:  looks like there are lots of opportunities for code sharing here with
the
> Diffuse benches.  You can do that in a followup patch if you like.

Done. I made some base class functionality to reduce code duplication.

[Stephen White, 2013-04-24 19:27:12]

LGTM.  Thanks!

[sugoi1, 2013-04-24 19:36:48]

Committed patchset #3 manually as r8846 (presubmit successful).