update memset16/32 inlining heuristics

update memset16/32 inlining heuristics

I spent some time looking at perf.skia.org and it looks like we can do better.

It is weird, weird, weird that on x86, we see three completely different behaviors:
  - x86 Android: inlining better for small N, custom better for large N;
  - Windows:     inlining better for large N, custom better for small N;
  - other x86:   inlining generally better

BUG=skia:4316, chromium:516426
Committed: https://skia.googlesource.com/skia/+/b68fa409fc00ce2f38e2a0fd6f9dc2379b372481

Summaries: https://perf.skia.org/#4179
All traces, log scale: https://perf.skia.org/#4180

TBR=reed@google.com
No public API changes.

Committed: https://skia.googlesource.com/skia/+/3d096654b910e52768d7335a0c2c7d7cd32c8bb7

[mtklein_C, 2015-09-20 21:51:44]

mtklein@chromium.org changed reviewers:
+ jcgregorio@google.com, reed@google.com

[mtklein_C, 2015-09-20 21:51:45]

Joe, any chance perf.skia.org + trybots is nearing beta?  This'd be one where
I'd be interested in seeing the effect on tests containing 'memset' on every
type of CPU bot.

[mtklein_C, 2015-09-20 21:53:39]

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

[commit-bot: I haz the power, 2015-09-20 21:53:42]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1357193002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1357193002/1

[commit-bot: I haz the power, 2015-09-20 21:58:32]

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

[commit-bot: I haz the power, 2015-09-20 21:58:33]

Dry run: This issue passed the CQ dry run.

[mtklein, 2015-09-20 22:00:57]

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

[mtklein, 2015-09-20 22:00:57]

Hmm.  I've had an idea.  I'm going to land this CL then revert it right after it
lands.  The bots will still backfill perf.skia.org, so I can see which direction
this all goes.

[mtklein, 2015-09-20 22:01:22]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2015-09-20 22:01:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1357193002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1357193002/1

[commit-bot: I haz the power, 2015-09-20 22:02:18]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/b68fa409fc00ce2f38e2a0fd6f9dc2379b372481

[mtklein, 2015-09-20 22:02:46]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/1358793002/ by mtklein@google.com.

The reason for reverting is: Who wants to land forever?.

[jcgregorio, 2015-09-23 20:50:45]

lgtm

[mtklein_C, 2015-09-29 17:29:32]

The CQ bit was checked by mtklein@chromium.org

[mtklein_C, 2015-09-29 17:29:32]

The patchset sent to the CQ was uploaded after l-g-t-m from
jcgregorio@google.com
Link to the patchset: https://codereview.chromium.org/1357193002/#ps20001
(title: "dont forget armv7, INLINE_IF")

[commit-bot: I haz the power, 2015-09-29 17:29:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1357193002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1357193002/20001

[commit-bot: I haz the power, 2015-09-29 17:39:03]

Committed patchset #2 (id:20001) as
https://skia.googlesource.com/skia/+/3d096654b910e52768d7335a0c2c7d7cd32c8bb7

[Noel Gordon, 2015-11-25 02:02:24]

noel@chromium.org changed reviewers:
+ noel@chromium.org

[Noel Gordon, 2015-11-25 02:02:25]

https://codereview.chromium.org/1357193002/diff/20001/include/core/SkUtils.h
File include/core/SkUtils.h (right):

https://codereview.chromium.org/1357193002/diff/20001/include/core/SkUtils.h#...
include/core/SkUtils.h:25: #define INLINE_IF(cond) if (cond) { while (count -->
0) { *buffer++ = value; } return; }
while (count --> 0)

I suppose it compiles, but maybe 

while (count-- > 0)

might be idiomatic here?