Add HardStopGradientBench_ScaleNumHardStops.cpp

Add HardStopGradientBench_ScaleNumHardStops.cpp


Rename HardStopGradientBench.cpp to HardStopGradientBench_ScaleNumColors.cpp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2178913003

Committed: https://skia.googlesource.com/skia/+/17e829794db9621879baf2e90856a72b17eaf7dc

[fmenozzi, 2016-07-25 15:37:07]

Description was changed from

==========
Add HardStopGradientBench_ScaleNumHardStops.cpp


Rename HardStopGradientBench.cpp to HardStopGradientBench_ScaleNumColors.cpp

BUG=skia:
==========

to

==========
Add HardStopGradientBench_ScaleNumHardStops.cpp


Rename HardStopGradientBench.cpp to HardStopGradientBench_ScaleNumColors.cpp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2178913003
==========

[fmenozzi, 2016-07-25 17:13:27]

fmenozzi@google.com changed reviewers:
+ tomhudson@google.com

[fmenozzi, 2016-07-25 17:13:28]

Rather than adding to the previous benchmark, I decided to add another and
rename the original to better reflect the differences between the two. So now,
instead of having a single HardStopGradientBench, we have
HardStopGradientBench_ScaleNumColors and
HardStopGradientBench_ScaleNumHardStops; the former scales the number of colors
in a gradient with a single hard stop, while the latter also scales the number
of hard stops.

If you feel like the specific numbers chosen for the latter are not the best,
feel free to say so; I chose the numbers that I did in order to (hopefully)
demonstrate that

1) The running time should scale roughly linearly in the number of hard stops
instead of the number of colors

2) Two gradients with the same number of hard stops should render in roughly the
same amount of time, regardless of the number of colors

ptal

[fmenozzi, 2016-07-26 12:59:06]

bump

[tomhudson, 2016-07-26 13:25:14]

https://codereview.chromium.org/2178913003/diff/1/bench/HardStopGradientBench...
File bench/HardStopGradientBench_ScaleNumColors.cpp (right):

https://codereview.chromium.org/2178913003/diff/1/bench/HardStopGradientBench...
bench/HardStopGradientBench_ScaleNumColors.cpp:53: * trigger the hard stop code.
Nit: runon sentence?

https://codereview.chromium.org/2178913003/diff/1/bench/HardStopGradientBench...
File bench/HardStopGradientBench_ScaleNumHardStops.cpp (right):

https://codereview.chromium.org/2178913003/diff/1/bench/HardStopGradientBench...
bench/HardStopGradientBench_ScaleNumHardStops.cpp:55: positions[k++] = 0.0f;
This is not right; it looks like you're creating one stop of multiplicity
fHardStopCount, so with a parameter of 1 you'd be creating a soft stop?

[fmenozzi, 2016-07-26 13:32:41]

Add Tom's suggestions, ptal

[tomhudson, 2016-07-26 13:42:28]

https://codereview.chromium.org/2178913003/diff/20001/bench/HardStopGradientB...
File bench/HardStopGradientBench_ScaleNumHardStops.cpp (right):

https://codereview.chromium.org/2178913003/diff/20001/bench/HardStopGradientB...
bench/HardStopGradientBench_ScaleNumHardStops.cpp:56: positions[k++] = 0.0f;
Setting a 2*k-multiplicity hard stop at 0 isn't the same as setting k
2-multiplicity hard stops at different coordinates, is it?

[fmenozzi, 2016-07-26 13:46:50]

Third time's the charm, ptal

[tomhudson, 2016-07-26 13:54:27]

lgtm

[fmenozzi, 2016-07-27 20:40:24]

Fixed problems related to color and position distributions in both the
benchmarks, ptal

[fmenozzi, 2016-07-28 14:06:08]

bumpity bumpity

[tomhudson, 2016-07-28 15:30:55]

lgtm

[fmenozzi, 2016-07-28 15:53:34]

The CQ bit was checked by fmenozzi@google.com

[commit-bot: I haz the power, 2016-07-28 15:53:40]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-28 16:44:51]

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

[commit-bot: I haz the power, 2016-07-28 16:44:52]

Try jobs failed on following builders:
  Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-Trybot on master.client.skia
(JOB_FAILED,
http://build.chromium.org/p/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2...)

[fmenozzi, 2016-07-28 16:54:49]

The CQ bit was checked by fmenozzi@google.com

[fmenozzi, 2016-07-28 16:54:50]

The patchset sent to the CQ was uploaded after l-g-t-m from tomhudson@google.com
Link to the patchset: https://codereview.chromium.org/2178913003/#ps80001
(title: "Fix off by one error in SkAssert")

[commit-bot: I haz the power, 2016-07-28 16:55:01]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-28 17:59:51]

Description was changed from

==========
Add HardStopGradientBench_ScaleNumHardStops.cpp


Rename HardStopGradientBench.cpp to HardStopGradientBench_ScaleNumColors.cpp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2178913003
==========

to

==========
Add HardStopGradientBench_ScaleNumHardStops.cpp


Rename HardStopGradientBench.cpp to HardStopGradientBench_ScaleNumColors.cpp

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2178913003

Committed:
https://skia.googlesource.com/skia/+/17e829794db9621879baf2e90856a72b17eaf7dc
==========

[commit-bot: I haz the power, 2016-07-28 17:59:52]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/17e829794db9621879baf2e90856a72b17eaf7dc