non-aa rects batch

non-aa rects batch

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/ee72dde696b35534465c14f6b4c5bfca44a2f63e

Committed: https://skia.googlesource.com/skia/+/58773334d7f4f27b11458e0ee77f60fe0ed939b9

[joshualitt, 2015-02-20 14:37:31]

joshualitt@google.com changed reviewers:
+ robertphillips@google.com

[joshualitt, 2015-02-20 14:47:02]

joshualitt@google.com changed reviewers:
+ joshualitt@google.com

[joshualitt, 2015-02-20 14:47:03]

ptal

[robertphillips, 2015-02-20 17:29:38]

lgtm

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
File src/gpu/GrInOrderDrawBuffer.cpp (right):

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:110: public:
That'a a lot of data!

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:114: SkRect fRect;
Does reordering the bools yield better packing?

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:284: if (!this->hasLocalRect() &&
this->usesLocalCoords()) {
Do we need a view matrix for each one then?

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:292: 
Ditto for local matrix?

[joshualitt, 2015-02-20 17:49:07]

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
File src/gpu/GrInOrderDrawBuffer.cpp (right):

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:114: SkRect fRect;
On 2015/02/20 17:29:38, robertphillips wrote:
> Does reordering the bools yield better packing?

I can try putting them at the end, that is a good idea.  I would think on most
platforms it will be tightly packed, but I very well could be wrong.

https://codereview.chromium.org/931293002/diff/80001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:284: if (!this->hasLocalRect() &&
this->usesLocalCoords()) {
On 2015/02/20 17:29:38, robertphillips wrote:
> Do we need a view matrix for each one then?

Well, I think the most common case by far is to not have explicit local coords,
so in reality we can batch most of the time across viewmatrix changes.  I could
make two batches though, if this ends up being a bottleneck.  Its is a ton of
data per geometry.  If the perf bots complain we can take a closer look at
better ways to optimize this.

[joshua.litt, 2015-02-20 22:36:22]

New patchsets have been uploaded after l-g-t-m from robertphillips@google.com

[joshualitt, 2015-02-23 14:58:38]

The CQ bit was checked by joshualitt@google.com

[commit-bot: I haz the power, 2015-02-23 14:59:29]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/931293002/90001

[commit-bot: I haz the power, 2015-02-23 15:16:13]

Committed patchset #6 (id:90001) as
https://skia.googlesource.com/skia/+/ee72dde696b35534465c14f6b4c5bfca44a2f63e

[joshualitt, 2015-02-23 15:34:25]

A revert of this CL (patchset #6 id:90001) has been created in
https://codereview.chromium.org/944743005/ by joshualitt@google.com.

The reason for reverting is: breaks angle.

[joshualitt, 2015-02-24 00:20:29]

The CQ bit was checked by joshualitt@google.com

[commit-bot: I haz the power, 2015-02-24 00:21:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/931293002/110001

[commit-bot: I haz the power, 2015-02-24 00:41:46]

Committed patchset #7 (id:110001) as
https://skia.googlesource.com/skia/+/58773334d7f4f27b11458e0ee77f60fe0ed939b9