Xfermode: SSE2 implementation of darken&lighten mode

Xfermode: SSE2 implementation of darken&lighten mode

With SSE2 optimization, performance of the related two benchmarks will
improve about 45% on desktop i7-3770. Here are the data:
before:
Xfermode_Lighten   8888:  cmsecs =     33.60   565:  cmsecs =     48.84
 Xfermode_Darken   8888:  cmsecs =     34.16   565:  cmsecs =     48.99
after:
Xfermode_Lighten   8888:  cmsecs =     18.71   565:  cmsecs =     25.41
 Xfermode_Darken   8888:  cmsecs =     18.39   565:  cmsecs =     25.40

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=14395

[qiankun, 2014-04-11 05:39:58]

PTAL.

[mtklein, 2014-04-24 17:55:11]

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
File src/opts/SkXfermode_opts_SSE2.cpp (right):

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
src/opts/SkXfermode_opts_SSE2.cpp:94: __m128i ret1 = _mm_add_epi32(sc, dc);
I'm not sure why the original code was written so oddly, but we can probably be
pretty sure that _mm_add_epi32(sc, dc) and _mm_add_epi32(dc, sc) are the same
value.  I figure it's worth saving that in its own register here rather than
recalculating?

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
src/opts/SkXfermode_opts_SSE2.cpp:127: __m128i ret1 = _mm_add_epi32(sc, dc);
Same here.

[qiankun, 2014-04-25 14:13:24]

updated the CL, PTAL.

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
File src/opts/SkXfermode_opts_SSE2.cpp (right):

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
src/opts/SkXfermode_opts_SSE2.cpp:94: __m128i ret1 = _mm_add_epi32(sc, dc);
On 2014/04/24 17:55:12, mtklein wrote:
> I'm not sure why the original code was written so oddly, but we can probably
be
> pretty sure that _mm_add_epi32(sc, dc) and _mm_add_epi32(dc, sc) are the same
> value.  I figure it's worth saving that in its own register here rather than
> recalculating?

Done.

https://codereview.chromium.org/234653002/diff/1/src/opts/SkXfermode_opts_SSE...
src/opts/SkXfermode_opts_SSE2.cpp:127: __m128i ret1 = _mm_add_epi32(sc, dc);
On 2014/04/24 17:55:12, mtklein wrote:
> Same here.

Done.

[mtklein, 2014-04-25 14:33:59]

lgtm

[qiankun, 2014-04-26 05:04:10]

The CQ bit was checked by qiankun.miao@intel.com

[commit-bot: I haz the power, 2014-04-26 05:04:41]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/qiankun.miao@intel.com/234653002/20001

[commit-bot: I haz the power, 2014-04-28 13:26:55]

Change committed as 14395