Update how we send draws to gpu backend to reduce state setting.

Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, we set the gpu state on gpu once per pipeline/prim proc draw batch, and we create the ProgramDescriptor only for the Cache/ProgramBuilder.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1806983002

Committed: https://skia.googlesource.com/skia/+/0e1853c89615d14d0d03c87c7e0c604e5285cc54

[egdaniel, 2016-03-16 17:50:27]

Description was changed from

==========
Update how we send draws to gpu backend to reduce state setting.

BUG=skia:
==========

to

==========
Update how we send draws to gpu backend to reduce state setting.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[egdaniel, 2016-03-16 19:52:57]

Description was changed from

==========
Update how we send draws to gpu backend to reduce state setting.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, and we
set the gpu state on gpu once per pipeline/prim proc draws.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[egdaniel, 2016-03-16 19:55:27]

Description was changed from

==========
Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, and we
set the gpu state on gpu once per pipeline/prim proc draws.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, we set
the gpu state on gpu once per pipeline/prim proc draw batch, and we create the
ProgramDescriptor only for the Cache/ProgramBuilder.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[egdaniel, 2016-03-16 19:55:34]

egdaniel@google.com changed reviewers:
+ bsalomon@google.com, ethannicholas@google.com, jvanverth@google.com

[egdaniel, 2016-03-16 19:55:35]

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:2932: if (!fHWPLSEnabled && plsState !=
Ethan do you know if all this PLS stuff needs to stay inside the inner loop or
can any of it move out at all?

[bsalomon, 2016-03-16 20:07:23]

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h
File src/gpu/GrGpu.h (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h#newcode381
src/gpu/GrGpu.h:381: struct DrawArgs {
Delete this?

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h#newcode402
src/gpu/GrGpu.h:402: int drawCount);
verticesCnt? Had to think for a bit to remember that it was the count for the
GrVertices array.

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:2932: if (!fHWPLSEnabled && plsState !=
On 2016/03/16 19:55:35, egdaniel wrote:
> Ethan do you know if all this PLS stuff needs to stay inside the inner loop or
> can any of it move out at all?

Seems like it could all move out since all we're doing here is un-instancing a
draw that the batch issued as a single logical draw.

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/vk/GrVkGpu.cpp
File src/gpu/vk/GrVkGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/vk/GrVkGpu.cpp#...
src/gpu/vk/GrVkGpu.cpp:1329: // Get GrGLProgramDesc
?

[ethannicholas, 2016-03-16 21:14:19]

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:2932: if (!fHWPLSEnabled && plsState !=
On 2016/03/16 20:07:22, bsalomon wrote:
> On 2016/03/16 19:55:35, egdaniel wrote:
> > Ethan do you know if all this PLS stuff needs to stay inside the inner loop
or
> > can any of it move out at all?
> 
> Seems like it could all move out since all we're doing here is un-instancing a
> draw that the batch issued as a single logical draw.

Agreed.

[egdaniel, 2016-03-17 15:09:36]

Updated to change GrVertices to GrMesh and put primitive type back on that
object

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h
File src/gpu/GrGpu.h (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h#newcode381
src/gpu/GrGpu.h:381: struct DrawArgs {
On 2016/03/16 20:07:22, bsalomon wrote:
> Delete this?

Done.

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/GrGpu.h#newcode402
src/gpu/GrGpu.h:402: int drawCount);
On 2016/03/16 20:07:22, bsalomon wrote:
> verticesCnt? Had to think for a bit to remember that it was the count for the
> GrVertices array.

changed to meshCount

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp
File src/gpu/gl/GrGLGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/gl/GrGLGpu.cpp#...
src/gpu/gl/GrGLGpu.cpp:2932: if (!fHWPLSEnabled && plsState !=
On 2016/03/16 21:14:19, ethannicholas wrote:
> On 2016/03/16 20:07:22, bsalomon wrote:
> > On 2016/03/16 19:55:35, egdaniel wrote:
> > > Ethan do you know if all this PLS stuff needs to stay inside the inner
loop
> or
> > > can any of it move out at all?
> > 
> > Seems like it could all move out since all we're doing here is un-instancing
a
> > draw that the batch issued as a single logical draw.
> 
> Agreed.

Done.

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/vk/GrVkGpu.cpp
File src/gpu/vk/GrVkGpu.cpp (right):

https://codereview.chromium.org/1806983002/diff/40001/src/gpu/vk/GrVkGpu.cpp#...
src/gpu/vk/GrVkGpu.cpp:1329: // Get GrGLProgramDesc
On 2016/03/16 20:07:22, bsalomon wrote:
> ?

Done.

[bsalomon, 2016-03-17 15:32:59]

lgtm

[egdaniel, 2016-03-17 18:12:28]

The CQ bit was checked by egdaniel@google.com

[commit-bot: I haz the power, 2016-03-17 18:12:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1806983002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1806983002/100001

[commit-bot: I haz the power, 2016-03-17 18:35:52]

Description was changed from

==========
Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, we set
the gpu state on gpu once per pipeline/prim proc draw batch, and we create the
ProgramDescriptor only for the Cache/ProgramBuilder.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
Update how we send draws to gpu backend to reduce state setting.

The main change here is that we pull primitive type off of the vertices, we set
the gpu state on gpu once per pipeline/prim proc draw batch, and we create the
ProgramDescriptor only for the Cache/ProgramBuilder.

BUG=skia:
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/0e1853c89615d14d0d03c87c7e0c604e5285cc54
==========

[commit-bot: I haz the power, 2016-03-17 18:35:53]

Committed patchset #6 (id:100001) as
https://skia.googlesource.com/skia/+/0e1853c89615d14d0d03c87c7e0c604e5285cc54