gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla; https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/0c8b1229552985d8ed976b8d47e66ff983afceac
Cr-Commit-Position: refs/heads/master@{#429240}

[dshwang, 2016-10-28 18:00:57]

Description was changed from

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug. I'll file this issue to Mesa bugzilla.

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One:                 25 FPS           22.5 FPS

BUG=535198
==========

to

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug. I'll file this issue to Mesa bugzilla.

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One:                 25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[inactive_dshwang_plz_cc_intel, 2016-10-28 18:01:24]

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

[dshwang, 2016-10-28 18:01:59]

Description was changed from

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug. I'll file this issue to Mesa bugzilla.

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One:                 25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug. I'll file this issue to Mesa bugzilla.

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):    25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-10-28 18:02:25]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[dshwang, 2016-10-28 18:10:46]

Description was changed from

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug. I'll file this issue to Mesa bugzilla.

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):    25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):    25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[dshwang, 2016-10-28 18:13:23]

dongseong.hwang@intel.com changed reviewers:
+ adrian.belgun@intel.com, piman@chromium.org

[dshwang, 2016-10-28 18:13:24]

piman@, could you review this strange issue?

adrian@, it may be what you mentioned earlier. I think it resolves your concern.

https://codereview.chromium.org/2460973002/diff/1/gpu/command_buffer/service/...
File
gpu/command_buffer/service/gles2_cmd_apply_framebuffer_attachment_cmaa_intel.cc
(right):

https://codereview.chromium.org/2460973002/diff/1/gpu/command_buffer/service/...
gpu/command_buffer/service/gles2_cmd_apply_framebuffer_attachment_cmaa_intel.cc:556:
const bool draw = true;
Mesa may will fix it or future Intel Chromebook may have different behavior.
So I remain dead code here. It makes us to test very quickly.

[dshwang, 2016-10-28 18:56:43]

Description was changed from

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glDrawArrays  glCopyTexSubImage2D
 4k fish on Chromebook Pixel 2015 (Broadwell):    22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                   23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):    25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-10-28 19:04:01]

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

[commit-bot: I haz the power, 2016-10-28 19:04:03]

Dry run: Try jobs failed on following builders:
  linux_chromium_chromeos_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[piman, 2016-10-28 21:51:12]

https://codereview.chromium.org/2460973002/diff/1/gpu/command_buffer/service/...
File
gpu/command_buffer/service/gles2_cmd_apply_framebuffer_attachment_cmaa_intel.cc
(right):

https://codereview.chromium.org/2460973002/diff/1/gpu/command_buffer/service/...
gpu/command_buffer/service/gles2_cmd_apply_framebuffer_attachment_cmaa_intel.cc:551:
void ApplyFramebufferAttachmentCMAAINTELResourceManager::CopyTexture(
We already have a facility to copy a texture using a draw, in
CopyTextureCHROMIUMResourceManager - that is used for CopyTextureCHROMIUM

Could we use that instead, to avoid duplicating functionality? We can pass the
CopyTextureCHROMIUMResourceManager to the
ApplyFramebufferAttachmentCMAAINTELResourceManager so that it can call into it.

[dshwang, 2016-11-01 17:09:03]

Description was changed from

==========
gpu, cmaa: optimize copy.

If input texture format is not RGBA8, CMAA copies the input texture to RGBA8.
Currently, CMAA copies texture using glCopyTexSubImage2D, but copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in 
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[dshwang, 2016-11-01 17:12:46]

piman@, could you review?

After landing https://codereview.chromium.org/2465963002/, I rebase it.
Interestingly, CopyTextureCHROMIUMResourceManager is fater than previous custom
shader.
for 4k aquarium on Pixel 2, this CL speed up 32.5 FPS -> 34.5 FPS
I cannot understand where the speed up if from.

[inactive_dshwang_plz_cc_intel, 2016-11-01 17:14:11]

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

[commit-bot: I haz the power, 2016-11-01 17:14:32]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[dshwang, 2016-11-01 17:33:23]

Patchset #2 (id:20001) has been deleted

[inactive_dshwang_plz_cc_intel, 2016-11-01 17:33:23]

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

[dshwang, 2016-11-01 17:33:39]

Patchset #2 (id:20001) has been deleted

[commit-bot: I haz the power, 2016-11-01 17:33:49]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-01 18:14:33]

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

[commit-bot: I haz the power, 2016-11-01 18:14:34]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  mac_optional_gpu_tests_rel on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_optional_gpu_...)

[piman, 2016-11-01 19:04:59]

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
File gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc (right):

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc:334: NOTREACHED();
I think this needs to be expanded to more formats. See test failures.

You can take a look at the logic we have in GLES2DecoderImpl::DoCopyTexImage2D
to check the API behavior.

Maybe here you can use TextureManager::ExtractFormatFromStorageFormat ?

[inactive_dshwang_plz_cc_intel, 2016-11-01 19:21:59]

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

[commit-bot: I haz the power, 2016-11-01 19:22:17]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[dshwang, 2016-11-01 19:30:06]

Thank you for review.

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
File gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc (right):

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc:334: NOTREACHED();
On 2016/11/01 19:04:59, piman wrote:
> I think this needs to be expanded to more formats. See test failures.
exactly, this DCHECK causes crash. Done.

> You can take a look at the logic we have in GLES2DecoderImpl::DoCopyTexImage2D
> to check the API behavior.

It's used by only below SourceFormatContainSupersetOfDestFormat, so I add
sufficient format to satisfy the below function.

> Maybe here you can use TextureManager::ExtractFormatFromStorageFormat ?

The function consider GL_COMPRESSED_RGBA_S3TC_DXT1_EXT as RGBA, which is
impossible to use glCopyTexSubImage2D.

I want to express
(UnsizedFormatFrom(source_internal_format) ==
UnsizedFormatFrom(dest_internal_format) && (source_internal_format !=
GL_BGRA_EXT && source_internal_format != GL_BGRA8_EXT) ||
(source_internal_format == GL_RGBA && dest_internal_format) == GL_RGB) ||
(source_internal_format == GL_RGBA8 && dest_internal_format) == GL_RGB) ||
(source_internal_format == GL_RGBA && dest_internal_format) == GL_RGB8) ||
(source_internal_format == GL_RGBA8 && dest_internal_format) == GL_RGB8);

However, it's too long so just make this helper function.

[inactive_dshwang_plz_cc_intel, 2016-11-01 19:34:51]

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

[dshwang, 2016-11-01 19:35:17]

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
File gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc (right):

https://codereview.chromium.org/2460973002/diff/40001/gpu/command_buffer/serv...
gpu/command_buffer/service/gles2_cmd_copy_texture_chromium.cc:334: NOTREACHED();
On 2016/11/01 19:30:05, dshwang wrote:
> On 2016/11/01 19:04:59, piman wrote:
> > I think this needs to be expanded to more formats. See test failures.
> exactly, this DCHECK causes crash. Done.
> 
> > You can take a look at the logic we have in
GLES2DecoderImpl::DoCopyTexImage2D
> > to check the API behavior.
> 
> It's used by only below SourceFormatContainSupersetOfDestFormat, so I add
> sufficient format to satisfy the below function.
> 
> > Maybe here you can use TextureManager::ExtractFormatFromStorageFormat ?
> 
> The function consider GL_COMPRESSED_RGBA_S3TC_DXT1_EXT as RGBA, which is
> impossible to use glCopyTexSubImage2D.
> 
> I want to express
> (UnsizedFormatFrom(source_internal_format) ==
> UnsizedFormatFrom(dest_internal_format) && (source_internal_format !=
> GL_BGRA_EXT && source_internal_format != GL_BGRA8_EXT) ||
> (source_internal_format == GL_RGBA && dest_internal_format) == GL_RGB) ||
> (source_internal_format == GL_RGBA8 && dest_internal_format) == GL_RGB) ||
> (source_internal_format == GL_RGBA && dest_internal_format) == GL_RGB8) ||
> (source_internal_format == GL_RGBA8 && dest_internal_format) == GL_RGB8);
> 
> However, it's too long so just make this helper function.

I gave up this non-trivial change which cause lots of webgl conformance test
failures on Mac. It's not needed in this CL. So remove it.

I cannot understand why mac fail. I'll figure out and submit another CL about
it.

[commit-bot: I haz the power, 2016-11-01 19:35:24]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[piman, 2016-11-01 19:38:56]

lgtm

[commit-bot: I haz the power, 2016-11-01 20:33:50]

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

[commit-bot: I haz the power, 2016-11-01 20:33:51]

Dry run: This issue passed the CQ dry run.

[dshwang, 2016-11-02 08:33:05]

The CQ bit was checked by dongseong.hwang@intel.com

[commit-bot: I haz the power, 2016-11-02 08:33:20]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-02 08:38:54]

Description was changed from

==========
gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in 
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in 
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-11-02 08:38:55]

Committed patchset #4 (id:80001)

[commit-bot: I haz the power, 2016-11-02 08:44:17]

Description was changed from

==========
gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in 
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
gpu, cmaa: reuse CopyTextureCHROMIUMResourceManager

In addition, prevent from using glCopyTexSubImage2D in
CopyTextureCHROMIUMResourceManager on Intel ChromeOS, because copy via
glDrawArrays is much faster in IA ChromeOS. However, glCopyTexSubImage2D is
faster in Android.

It looks Mesa bug, so I filed this issue to Mesa bugzilla;
https://bugs.freedesktop.org/show_bug.cgi?id=98478

Perf data >
 test site: http://webglsamples.org/aquarium/aquarium.html
                                               glCopyTexSubImage2D glDrawArrays
 4k fish on Chromebook Pixel 2015 (Broadwell):      22 FPS           32.6 FPS
 4k fish on Ubuntu and Haswell:                     23 FPS           30.9 FPS
 500 fish on Android OnePlus One (Adreno 330):      25 FPS           22.5 FPS

BUG=535198
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/0c8b1229552985d8ed976b8d47e66ff983afceac
Cr-Commit-Position: refs/heads/master@{#429240}
==========

[commit-bot: I haz the power, 2016-11-02 08:44:19]

Patchset 4 (id:??) landed as
https://crrev.com/0c8b1229552985d8ed976b8d47e66ff983afceac
Cr-Commit-Position: refs/heads/master@{#429240}