Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and I422ToARGB1555Row_MSA functions

Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and I422ToARGB1555Row_MSA functions

R=fbarchard@google.com
BUG=libyuv:634
Performance Gain (vs C vectorized)
I422ToRGB565Row_MSA             : ~1.5x
I422ToRGB565Row_Any_MSA         : ~1.5x
I422ToARGB4444Row_MSA           : ~1.4x
I422ToARGB4444Row_Any_MSA       : ~1.4x
I422ToARGB1555Row_MSA           : ~1.4x
I422ToARGB1555Row_Any_MSA       : ~1.4x

Performance Gain (vs C non-vectorized)
I422ToRGB565Row_MSA             : ~6.8x
I422ToRGB565Row_Any_MSA         : ~6.8x
I422ToARGB4444Row_MSA           : ~6.6x
I422ToARGB4444Row_Any_MSA       : ~6.6x
I422ToARGB1555Row_MSA           : ~6.6x
I422ToARGB1555Row_Any_MSA       : ~6.6x

Committed: https://chromium.googlesource.com/libyuv/libyuv/+/10ce829bad9b1d9908874b8aa9f3ba8d04523282

[manojkumar.bhosale, 2016-10-27 12:04:18]

Description was changed from

==========
Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and
I422ToARGB1555Row_MSA functions

R=fbarchard@google.com
BUG=libyuv:634
==========

to

==========
Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and
I422ToARGB1555Row_MSA functions

R=fbarchard@google.com
BUG=libyuv:634

Performance Gain (vs C vectorized)
I422ToRGB565Row_MSA             : ~1.5x
I422ToRGB565Row_Any_MSA         : ~1.5x
I422ToARGB4444Row_MSA           : ~1.4x
I422ToARGB4444Row_Any_MSA       : ~1.4x
I422ToARGB1555Row_MSA           : ~1.4x
I422ToARGB1555Row_Any_MSA       : ~1.4x

Performance Gain (vs C non-vectorized)
I422ToRGB565Row_MSA             : ~6.8x
I422ToRGB565Row_Any_MSA         : ~6.8x
I422ToARGB4444Row_MSA           : ~6.6x
I422ToARGB4444Row_Any_MSA       : ~6.6x
I422ToARGB1555Row_MSA           : ~6.6x
I422ToARGB1555Row_Any_MSA       : ~6.6x
==========

[manojkumar.bhosale, 2016-10-27 12:04:18]

manojkumar.bhosale@imgtec.com changed reviewers:
+ kjellander@chromium.org

[manojkumar.bhosale, 2016-10-27 12:04:48]

Updated with reviewers, cc list and performance gain numbers

[fbarchard1, 2016-10-27 17:37:15]

lgtm

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc
File source/row_msa.cc (right):

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc#newcode23
source/row_msa.cc:23: #define FILLYUVTORGBCONSTS(yuvconst, ub, vr, ug, vg, bb,
bg, br, yg) {  \
suggest using the same name as row_gcc: YUVTORGB_SETUP

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc#newcode341
source/row_msa.cc:341: I422TORGB(src0, src1, src2, vec_ub, vec_vr, vec_ug,
vec_vg,
Suggest renaming I422ToRGB to YUVTORGB for consistency with row_gcc.cc

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc#newcode376
source/row_msa.cc:376: reg1 = (v8u16) __msa_srai_h(vec1, 4);
instead of shift right and then left, consider in place AND

[fbarchard1, 2016-10-27 17:47:39]

Description was changed from

==========
Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and
I422ToARGB1555Row_MSA functions

R=fbarchard@google.com
BUG=libyuv:634

Performance Gain (vs C vectorized)
I422ToRGB565Row_MSA             : ~1.5x
I422ToRGB565Row_Any_MSA         : ~1.5x
I422ToARGB4444Row_MSA           : ~1.4x
I422ToARGB4444Row_Any_MSA       : ~1.4x
I422ToARGB1555Row_MSA           : ~1.4x
I422ToARGB1555Row_Any_MSA       : ~1.4x

Performance Gain (vs C non-vectorized)
I422ToRGB565Row_MSA             : ~6.8x
I422ToRGB565Row_Any_MSA         : ~6.8x
I422ToARGB4444Row_MSA           : ~6.6x
I422ToARGB4444Row_Any_MSA       : ~6.6x
I422ToARGB1555Row_MSA           : ~6.6x
I422ToARGB1555Row_Any_MSA       : ~6.6x
==========

to

==========
Add MSA optimized I422ToRGB565Row_MSA, I422ToARGB4444Row_MSA and
I422ToARGB1555Row_MSA functions

R=fbarchard@google.com
BUG=libyuv:634

Performance Gain (vs C vectorized)
I422ToRGB565Row_MSA             : ~1.5x
I422ToRGB565Row_Any_MSA         : ~1.5x
I422ToARGB4444Row_MSA           : ~1.4x
I422ToARGB4444Row_Any_MSA       : ~1.4x
I422ToARGB1555Row_MSA           : ~1.4x
I422ToARGB1555Row_Any_MSA       : ~1.4x

Performance Gain (vs C non-vectorized)
I422ToRGB565Row_MSA             : ~6.8x
I422ToRGB565Row_Any_MSA         : ~6.8x
I422ToARGB4444Row_MSA           : ~6.6x
I422ToARGB4444Row_Any_MSA       : ~6.6x
I422ToARGB1555Row_MSA           : ~6.6x
I422ToARGB1555Row_Any_MSA       : ~6.6x

Committed:
https://chromium.googlesource.com/libyuv/libyuv/+/10ce829bad9b1d9908874b8aa9f...
==========

[fbarchard1, 2016-10-27 17:47:40]

Committed patchset #1 (id:1) manually as
10ce829bad9b1d9908874b8aa9f3ba8d04523282 (presubmit successful).

[fbarchard1, 2016-10-27 18:22:55]

rebased and landed

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc
File source/row_msa.cc (right):

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc#newcode23
source/row_msa.cc:23: #define FILLYUVTORGBCONSTS(yuvconst, ub, vr, ug, vg, bb,
bg, br, yg) {  \
On 2016/10/27 17:37:14, fbarchard1 wrote:
> suggest using the same name as row_gcc: YUVTORGB_SETUP

Done.

https://codereview.chromium.org/2445343007/diff/1/source/row_msa.cc#newcode341
source/row_msa.cc:341: I422TORGB(src0, src1, src2, vec_ub, vec_vr, vec_ug,
vec_vg,
On 2016/10/27 17:37:15, fbarchard1 wrote:
> Suggest renaming I422ToRGB to YUVTORGB for consistency with row_gcc.cc

Done.