optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

Before:
   8/8   MB	1	9.09ms	10.8ms	10.3ms	11.5ms	9%	█▆▆▁▁▁▂▅█▆	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8   MB	1	10.2ms	12.1ms	11.7ms	13.2ms	9%	▅▇▁▂▁▄█▆▅▆	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

After:
   8/8   MB	1	1.6ms	1.68ms	1.73ms	2.17ms	10%	▄▄█▁▃▁▁▂▁▁	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8   MB	1	3.13ms	3.62ms	3.97ms	5.81ms	21%	█▃▁▂▆▂▂▂▃▂	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1664713003

Committed: https://skia.googlesource.com/skia/+/6dfc754360234f2da7103e13e39146d8f702e8ef

[reed1, 2016-02-03 17:31:02]

Description was changed from

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

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

to

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[reed1, 2016-02-03 17:32:10]

Description was changed from

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

Before:
   8/8  
MB	1	9.09ms	10.8ms	10.3ms	11.5ms	9%	█▆▆▁▁▁▂▅█▆	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	10.2ms	12.1ms	11.7ms	13.2ms	9%	▅▇▁▂▁▄█▆▅▆	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

After:
   8/8  
MB	1	1.6ms	1.68ms	1.73ms	2.17ms	10%	▄▄█▁▃▁▁▂▁▁	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	3.13ms	3.62ms	3.97ms	5.81ms	21%	█▃▁▂▆▂▂▂▃▂	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[reed1, 2016-02-03 17:34:46]

reed@google.com changed reviewers:
+ herb@google.com, msarett@google.com, mtklein@google.com

[reed1, 2016-02-03 17:34:46]

For this unroll, I needed to do the following:

4 bytes of coverage (for 4 pixels)
into
[ aa[0], aa[0], aa[0], aa[0] ]
[ aa[1], aa[1], aa[1], aa[1] ]
[ aa[2], aa[2], aa[2], aa[2] ]
[ aa[3], aa[3], aa[3], aa[3] ]

Not sure if there's a more concise/efficient way to do this.

[reed1, 2016-02-03 17:36:02]

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

[commit-bot: I haz the power, 2016-02-03 17:36:12]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1664713003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1664713003/20001

[mtklein, 2016-02-03 17:39:24]

On 2016/02/03 17:34:46, reed1 wrote:
> For this unroll, I needed to do the following:
> 
> 4 bytes of coverage (for 4 pixels)
> into
> [ aa[0], aa[0], aa[0], aa[0] ]
> [ aa[1], aa[1], aa[1], aa[1] ]
> [ aa[2], aa[2], aa[2], aa[2] ]
> [ aa[3], aa[3], aa[3], aa[3] ]
> 
> Not sure if there's a more concise/efficient way to do this.

As you've written it is the fastest way I know how to do it today.


lgtm

[reed1, 2016-02-03 17:40:09]

I so love Sk4f_ToBytes -- 3x speedup!

[mtklein, 2016-02-03 17:40:11]

The linear speedup is >4x.  Do you believe that?

[reed1, 2016-02-03 17:41:56]

old version forced src and dst to be unit instead of keeping them 255 biased.
That could be part of the win too

[mtklein, 2016-02-03 17:43:17]

On 2016/02/03 17:41:56, reed1 wrote:
> old version forced src and dst to be unit instead of keeping them 255 biased.
> That could be part of the win too

Ah, of course.

[reed1, 2016-02-03 18:07:52]

The CQ bit was unchecked by reed@google.com

[reed1, 2016-02-03 18:07:56]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2016-02-03 18:08:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1664713003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1664713003/20001

[commit-bot: I haz the power, 2016-02-03 18:09:25]

Description was changed from

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

Before:
   8/8  
MB	1	9.09ms	10.8ms	10.3ms	11.5ms	9%	█▆▆▁▁▁▂▅█▆	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	10.2ms	12.1ms	11.7ms	13.2ms	9%	▅▇▁▂▁▄█▆▅▆	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

After:
   8/8  
MB	1	1.6ms	1.68ms	1.73ms	2.17ms	10%	▄▄█▁▃▁▁▂▁▁	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	3.13ms	3.62ms	3.97ms	5.81ms	21%	█▃▁▂▆▂▂▂▃▂	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
optimize src mode (opaque src in srcover) singleton with aa (e.g. a8 text mask)

Before:
   8/8  
MB	1	9.09ms	10.8ms	10.3ms	11.5ms	9%	█▆▆▁▁▁▂▅█▆	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	10.2ms	12.1ms	11.7ms	13.2ms	9%	▅▇▁▂▁▄█▆▅▆	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

After:
   8/8  
MB	1	1.6ms	1.68ms	1.73ms	2.17ms	10%	▄▄█▁▃▁▁▂▁▁	nonrendering	xfer4f_srcover_aa_1_opaque_linear
   8/8  
MB	1	3.13ms	3.62ms	3.97ms	5.81ms	21%	█▃▁▂▆▂▂▂▃▂	nonrendering	xfer4f_srcover_aa_1_opaque_srgb

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/6dfc754360234f2da7103e13e39146d8f702e8ef
==========

[commit-bot: I haz the power, 2016-02-03 18:09:25]

Committed patchset #2 (id:20001) as
https://skia.googlesource.com/skia/+/6dfc754360234f2da7103e13e39146d8f702e8ef