ARM optimization for S32_Opaque_D32_nofilter_DX

We extract a sub-function from MAKENAME(_nofilter_DX), only dealing with reading one index array, indexing into src array and output to dst array.

Because of the scatter-gather nature, we can not do much burst/batch reading/writing to improve the performance.

We tried Neon vector instructions.  We also tried the hand optimize the compiler generated assembly (non-neon) code.  The latter seems to have better gain.  About 6% improvements, not much though...

[XinQi, 2010-06-01 17:05:07]

Update license information in header

[agl, 2010-06-03 21:39:42]

(If there are any other reviews that you're waiting on, please ping them.
Cheers.)

http://codereview.appspot.com/1127042/diff/2001/3002
File core/SkBitmapProcState.cpp (right):

http://codereview.appspot.com/1127042/diff/2001/3002#newcode90
core/SkBitmapProcState.cpp:90: #if defined(__ARM_HAVE_NEON) &&
defined(SK_CPU_LENDIAN)
add a !defined(SK_TEMPLATE_USE_NEON) clause here.

http://codereview.appspot.com/1127042/diff/2001/3002#newcode91
core/SkBitmapProcState.cpp:91: #define SK_TEMPLATE_USE_NEON
in this block, also #define SkBitmapProcState_TEMPLATE_USE_NEON

http://codereview.appspot.com/1127042/diff/2001/3002#newcode94
core/SkBitmapProcState.cpp:94: #if defined(__ARM_HAVE_NEON) &&
defined(SK_CPU_LENDIAN)
Change this condition to defined(SkBitmapProcState_TEMPLATE_USE_NEON)

http://codereview.appspot.com/1127042/diff/2001/3003
File core/SkBitmapProcState_sample.h (right):

http://codereview.appspot.com/1127042/diff/2001/3003#newcode62
core/SkBitmapProcState_sample.h:62: extern "C" void
S32_Opaque_D32_nofilter_DX_gether_neon(SkPMColor* SK_RESTRICT colors,
typically a newline would be inserted after the "(" and the other lines would be
indented four spaces. No big deal either way.

http://codereview.appspot.com/1127042/diff/2001/3003#newcode62
core/SkBitmapProcState_sample.h:62: extern "C" void
S32_Opaque_D32_nofilter_DX_gether_neon(SkPMColor* SK_RESTRICT colors,
I assume that "gether" means something? It's not a typo of "gather"?

http://codereview.appspot.com/1127042/diff/2001/3003#newcode117
core/SkBitmapProcState_sample.h:117: #endif //#ifdef USE_NEON
s/USE_NEON/!SK_TEMPLATE_USE_NEON/  ?

http://codereview.appspot.com/1127042/diff/2001/3001
File opts/S32_Opaque_D32_nofilter_DX_gether.S (right):

http://codereview.appspot.com/1127042/diff/2001/3001#newcode17
opts/S32_Opaque_D32_nofilter_DX_gether.S:17: .text
I can't read ARM, but please write a whole bunch of comment at the top here
saying what this function is doing, what arguments and outputs it has etc.

[XinQi, 2010-06-15 20:06:47]

Update patch upon comments

[agl, 2010-06-16 20:04:31]

r579

[agl, 2010-06-18 15:34:53]

On Wed, Jun 16, 2010 at 4:04 PM,  <agl@chromium.org> wrote:
> r579

Reverted. Broke the build.


AGL