Faster 4f gradient premul path

Faster 4f gradient premul path

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002

Committed: https://skia.googlesource.com/skia/+/3a2e45a6ed50c07cb1a710d3f7b74be796e61251

[f(malita), 2016-10-14 14:33:55]

Description was changed from

==========
Faster 4f gradient premul

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
==========

to

==========
Faster 4f gradient premul

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002
==========

[f(malita), 2016-10-14 14:36:43]

before:

  17/17 
MB	1	880µs	914µs	898µs	920µs	2%	██▇▇▇▁▁▁▁▁	8888	gradient_linear_mirror_3color_4f
  17/17 
MB	1	1.02ms	1.04ms	1.04ms	1.07ms	2%	▇▇▇█▄▁▁▁▁▁	8888	gradient_linear_mirror_hicolor_4f
  17/17 
MB	1	1.03ms	1.03ms	1.03ms	1.05ms	1%	▂▁█▁▁▁▁▁▁▁	8888	gradient_linear_mirror_4f
  17/17 
MB	1	840µs	844µs	844µs	847µs	0%	█▆▇▄▁▄▅▄▄▆	8888	gradient_linear_repeat_3color_4f
  17/17 
MB	1	1.01ms	1.02ms	1.02ms	1.02ms	0%	▅▃▆▂█▂▁▃▄▃	8888	gradient_linear_repeat_hicolor_4f
  17/17 
MB	1	1.03ms	1.03ms	1.03ms	1.03ms	0%	▅▃▆▃▃▃▂▁█▂	8888	gradient_linear_repeat_4f
  17/17 
MB	1	836µs	841µs	841µs	851µs	0%	▅▄█▄▃▃▃▃▁▂	8888	gradient_linear_clamp_3color_4f
  17/17 
MB	1	1ms	1ms	1ms	1.01ms	0%	▄█▅▁▂▁▂▄▂▂	8888	gradient_linear_clamp_hicolor_4f
  17/17 
MB	1	1.03ms	1.03ms	1.03ms	1.03ms	0%	▅▅▆▁▁▅█▅▄▄	8888	gradient_linear_clamp_4f

after:

  17/17 
MB	2	480µs	500µs	499µs	507µs	1%	▇▆▆▆▆▆▆▆█▁	8888	gradient_linear_mirror_3color_4f
  17/17 
MB	2	541µs	565µs	561µs	568µs	2%	▇▇▇▇█▇▇█▁▁	8888	gradient_linear_mirror_hicolor_4f
  17/17 
MB	2	489µs	490µs	491µs	498µs	1%	▁▂▁▁▂▁▁█▁▁	8888	gradient_linear_mirror_4f
  17/17 
MB	2	459µs	461µs	461µs	464µs	0%	█▆▆▄▄▂▁▃▃▃	8888	gradient_linear_repeat_3color_4f
  17/17 
MB	2	533µs	534µs	535µs	536µs	0%	██▄▄▄▄▂▄▁▁	8888	gradient_linear_repeat_hicolor_4f
  17/17 
MB	2	490µs	491µs	491µs	493µs	0%	▅▃▆▃▃█▁▂▁▂	8888	gradient_linear_repeat_4f
  17/17 
MB	2	459µs	461µs	460µs	462µs	0%	█▆▁▂▄▅▅▄▄▅	8888	gradient_linear_clamp_3color_4f
  17/17 
MB	2	525µs	525µs	525µs	527µs	0%	█▁▂▂▂▂▂▄▂▂	8888	gradient_linear_clamp_hicolor_4f
  17/17 
MB	2	486µs	488µs	488µs	490µs	0%	█▁▆▄▅▅▄▅▆▅	8888	gradient_linear_clamp_4f

[f(malita), 2016-10-14 14:36:57]

Description was changed from

==========
Faster 4f gradient premul

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002
==========

to

==========
Faster 4f gradient premul path

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002
==========

[f(malita), 2016-10-14 14:37:59]

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

[commit-bot: I haz the power, 2016-10-14 14:38:06]

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

[reed1, 2016-10-14 14:58:43]

lgtm

[commit-bot: I haz the power, 2016-10-14 15:13:58]

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

[commit-bot: I haz the power, 2016-10-14 15:13:59]

Dry run: This issue passed the CQ dry run.

[f(malita), 2016-10-14 15:17:20]

The CQ bit was checked by fmalita@chromium.org

[commit-bot: I haz the power, 2016-10-14 15:17:29]

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, 2016-10-14 15:18:26]

Description was changed from

==========
Faster 4f gradient premul path

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002
==========

to

==========
Faster 4f gradient premul path

Similar to https://codereview.chromium.org/2409583003/, perform the
premul in 4f.  It turns out it's even faster to avoid the 255 load
multiplication in this case.

Also includes some template plumbing because DstTraits<>::load now needs
to be premul-aware (previously it wasn't).

R=reed@google.com
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2416233002

Committed:
https://skia.googlesource.com/skia/+/3a2e45a6ed50c07cb1a710d3f7b74be796e61251
==========

[commit-bot: I haz the power, 2016-10-14 15:18:27]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/3a2e45a6ed50c07cb1a710d3f7b74be796e61251