|
5093ba67
|
2019-05-14T17:36:36
|
|
Implement resource management for GL_EXT_semaphore
This implements glGenSemaphoresEXT, glDeleteSemaphoresEXT, and
glIsSemaphoreEXT. It's not possible to do anything useful with them yet.
Bug: angleproject:3289
Change-Id: I20ad90dbcd3fc573a4650c8531d6e1b8ccf7ca9b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1623811
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
0a1eeb80
|
2019-05-13T13:53:18
|
|
Call robust resource init before object sync.
This should let the Vulkan back-end only need to sync the Image data
once before we use them.
Bug: angleproject:3427
Change-Id: I8c6e8794e861f855bddbf651997351e8415e3479
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1602912
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
3dbe480b
|
2019-05-06T15:59:07
|
|
Validating draw buffers now also considers color mask settings
It will always return true if all 4 channels of color mask are set
to false.
This should only apply to WebGL. Will fix this in a later patch.
Bug: chromium:958374
Change-Id: I46befaf3ae1b63027dfbb309ac32724c616025d3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1594629
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
fb201c5e
|
2019-04-03T14:57:35
|
|
Implement resource management for GL_EXT_memory_object
This implements glCreateMemoryObjectsEXT, glDeleteMemoryObjectsEXT, and
glIsMemoryObjectEXT. It's not possible to do anything useful with them
yet.
Bug: angleproject:3289
Change-Id: I8882b657e9de564b5f97f8dea87838f67b1928f8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1552025
Commit-Queue: Michael Spang <spang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
d02c17f9
|
2019-03-28T16:49:39
|
|
Mark vertex attributes dirty during context switches
gl::State::mDirtyCurrentValues is a second tier of
gl::State::DIRTY_BIT_CURRENT_VALUES bit.
Marking mDirtyCurrentValues dirty during context switches
treats them in the same way as other gl::State::mDirtyBits are treated.
This has regressed in
https://chromium-review.googlesource.com/c/angle/angle/+/722423
where individual current values bits were squashed into one.
Bug: chromium:920033
Change-Id: I778515f11c975f8d88bd2d8c08c49160fd011497
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1545009
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
f7f15ac2
|
2019-03-27T12:56:51
|
|
Fix deleting a buffer not updating VAO validation.
Deleting a buffer that is bound to a VAO should act as if
the application unbound the buffer. Unbinding the buffer
should update relevant validation caches. But we were
missing the logic that updates the validation caches.
This CL adds the necessary cache updates. It does not include a
regression test. The test was causing an unrelated regression that is
going to be a bit longer. It should not block this fix.
Bug: chromium:943538
Change-Id: Ib073cd07a230ca073a5b14bc054e961158a0097d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1536491
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
1f750d1e
|
2019-01-16T13:27:16
|
|
Even more inlining and micro-optimization.
Improves perf by about 3-4% on the Vulkan VBO state change test.
Bug: angleproject:3014
Change-Id: If4415e503cd883cf7de6387bf1ebbca0e5fc994e
Reviewed-on: https://chromium-review.googlesource.com/c/1393907
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
|
|
465d6090
|
2019-01-02T16:21:18
|
|
Add GL_ANGLE_provoking_vertex on D3D11 and GL.
This extension is a subset of GL_ARB_provoking_vertex without the
QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION query.
Bug: angleproject:2829
Change-Id: I907a4d16b7b13d3bbfb948842091eedd7b6a8b77
Reviewed-on: https://chromium-review.googlesource.com/c/1410289
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
6e18a238
|
2019-01-16T13:27:14
|
|
Optimize more front-end VertexArray binding.
Improves perf slightly (1-2%) in the Vulkan VBO state change test.
Bug: angleproject:3014
Change-Id: Ia8082b5b3f5e847a6b2775e896893fa8d38c1afd
Reviewed-on: https://chromium-review.googlesource.com/c/1393904
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
dd34b3b9
|
2019-01-16T09:59:54
|
|
Pack VertexAttribType enum.
This improves performance slightly in vertex array format checks.
Instead of needing to switch on GLenum values we can use packed arrays
and tables to determine the values we need.
Does not significantly affect performance but will enable future work.
Bug: angleproject:3074
Change-Id: I6f4821a463e9b41fe3f8c8967eb3ed4c1d6b84be
Reviewed-on: https://chromium-review.googlesource.com/c/1393903
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
dbc605ce
|
2019-01-04T16:39:14
|
|
Vulkan: Optimize VBO state changes.
Also has some minor optimizations for the front-end.
12% improvement on the Vulkan VBO change test.
Bug: angleproject:3014
Change-Id: I38e1a8194edfc14bfe57424be348cb9688e928f4
Reviewed-on: https://chromium-review.googlesource.com/c/1369286
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
e4109f27
|
2018-12-13T16:25:53
|
|
WebGL: validate texture format matches sampler type
WebGL requires that drawing produces INVALID_OPERATION if a texture's
format doesn't match the sampler type it is bound to. This is a little
confusing because samplers have two attributes that could be called
"type": addressing mode (2D/3D/Cube), and component format
(float/signed/unsigned/shadow). ANGLE already handled checking the
addressing mode; this change adds checking for the component format.
Fixes WebGL conformance test
conformance2/uniforms/incompatible-texture-type-for-sampler.html
Bug: chromium:809237
Change-Id: I52ebfecd92625e3ee10274cb5f548d7e53de72dd
Reviewed-on: https://chromium-review.googlesource.com/c/1377611
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: James Darpinian <jdarpinian@chromium.org>
|
|
c3dc5d48
|
2018-12-30T12:12:04
|
|
Merge gl::Context and gl::ContextState.
This reduces the number of indrections when accessing the Extensions
or Caps structures. It will provide a small speed-up to some methods.
It also cleans up the code.
Bug: angleproject:2966
Change-Id: Idddac70758c42c1c2b75c885d0cacc8a5c458685
Reviewed-on: https://chromium-review.googlesource.com/c/1392391
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Markus Tavenrath <matavenrath@nvidia.com>
|
|
3f01f53c
|
2018-12-11T15:13:51
|
|
ES31: Add state change tests for image
This change adds dirty bit state change tests for image textures.
Meanwhile, syncImages related codes are removed since syncTextures will do
all texture states sync.
Bug: angleproject:3015
Change-Id: I9b299c86af1d589e72c08c5d7c55ac74cc7833aa
Reviewed-on: https://chromium-review.googlesource.com/c/1390596
Commit-Queue: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
91002266
|
2018-12-12T16:05:24
|
|
Use angle::Result and pass a context pointer to TransformFeedbackImpl
BUG=angleproject:3020
Change-Id: Ib0877dd33f9a8c6ea57976642f1b904258cb6a86
Reviewed-on: https://chromium-review.googlesource.com/c/1374273
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
7c985f5c
|
2018-11-29T18:16:17
|
|
Make angle::Result an enum.
This moves away from a class type to a value type. This should improve
performance when using angle::Result as a return value. Previously the
generated code would return a pointer instead of a value.
Improves performance in the most targeted microbenchmark by 10%. In
more realistic scanarios it will have a smaller improvement. Also
simplifies the class implementation and usage.
Includes some unrelated code generation changes.
Bug: angleproject:2491
Change-Id: Ifcf86870bf1c00a2f73c39ea6e4f05ca705050aa
Reviewed-on: https://chromium-review.googlesource.com/c/1356139
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
132d15c2
|
2018-11-30T15:25:38
|
|
Clear uncleared textures as a dirty object.
This removes the check for robust resource init from prepareForDraw.
It should improve performance more when WebGL compatiblity is enabled.
Also inlines Context::syncDirtyBits and Context::prepareForDraw.
Reduces CPU overhead in all draw and dispatch calls up to 5%.
Bug: angleproject:2966
Change-Id: I7e9009a501c0331a6140bc78678aef27e38c3df2
Reviewed-on: https://chromium-review.googlesource.com/c/1347453
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
e3e680ca
|
2018-12-03T17:49:08
|
|
Remove State::syncProgramTextures.
Removes the concept of the program textures dirty object. Instead we
use a set of dirty bits to represent dirty texture samples. We mark
certain textures dirty and update state structures whenever there is
a new Texture/Program/Sampler bound, or when Texture/Program/Sampler
state changes.
This is in preparation for making clearing the uncleared active
textures into a dirty bit as well.
Also includes new dirty bit handling for texture image units. These are
a GLES 3.1 feature.
Bug: angleproject:2966
Change-Id: Ibb8619dd2669bb39fdbcd75e3685be9a8aeeee91
Reviewed-on: https://chromium-review.googlesource.com/c/1346649
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
ef9fcd91
|
2018-11-28T14:03:59
|
|
Clear draw attachments as a dirty object.
This allows us to skip 1/2 of the robust resource init check in
Context::setupDraw. The plan is to remove the other half in a follow-up
CL.
Most of the work of this CL was already handled. We just need to add
the right dirty object mask to the Context's draw dirty objects list.
We can mask out this check when robust resource is not enabled.
Bug: angleproject:2966
Change-Id: I97ec2497c95e5cdf52988e1ce85f7602206935f0
Reviewed-on: https://chromium-review.googlesource.com/c/1343140
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
|
|
f7f8c518
|
2018-11-18T15:56:45
|
|
Optimize State::syncDirtyObjects
Create a function table for State::syncDirtyObjects to avoid the
costly switch and inline calls along the Context::syncDirtyObjects
calling hierarchy.
Bug: angleproject:2975
Change-Id: I1ec797452af41bc767578e4017c8eccb7d83628b
Reviewed-on: https://chromium-review.googlesource.com/c/1340222
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
b980c563
|
2018-11-27T11:34:27
|
|
Reformat all cpp and h files.
This applies git cl format --full to all ANGLE sources.
Bug: angleproject:2986
Change-Id: Ib504e618c1589332a37e97696cdc3515d739308f
Reviewed-on: https://chromium-review.googlesource.com/c/1351367
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
774fe563
|
2018-11-18T14:14:49
|
|
Make all single line getter functions in State class inline functions.
Having getter functions in the State class non-inline results in more generated code to call the function than it'd take to just fetch the value from the class. For those cases inline the getter functions to reduce calling overhead and binary size.
Bug: angleproject:2973
Change-Id: Iddd14fd836ee89de69cdabfd58b95bcedc7e9e4e
Reviewed-on: https://chromium-review.googlesource.com/c/1340220
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
526392dd
|
2018-11-16T09:35:14
|
|
Use angle::Result in front-end (Part 9)
This removes the ANGLE_TRY_HANDLE macro. Also the internal uses of
gl::ErrorStreamBase. There are remaining uses in the validation code.
Further progress will be blocked on removing egl::Error and the use
of gl::Error in the validation layer.
Also reduces binary size by up to 4k.
Bug: angleproject:2491
Change-Id: I3e0481f99738f9f24256c10e73f3efcce9826a35
Reviewed-on: https://chromium-review.googlesource.com/c/1334427
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
|
|
9e63a813
|
2018-11-16T12:30:04
|
|
Revert "Minor inlining optimization to Context draw calls."
This reverts commit d9ee8bf9620748a65df8f97ae89879eef676caff.
Reason for revert: Slightly regresses tests that have dirty objects.
See https://chromeperf.appspot.com/report?sid=abdb797da7149beffa29dc8a0869c15d96794749eff50bcb36c07d85fe509096
Original change's description:
> Minor inlining optimization to Context draw calls.
>
> Reduces draw call overhead by up to 3%.
>
> Bug: angleproject:2966
> Change-Id: Ie7ddb61b905fefe59a06a1528f0a3fde4accaf74
> Reviewed-on: https://chromium-review.googlesource.com/c/1333608
> Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
> Reviewed-by: Geoff Lang <geofflang@chromium.org>
> Commit-Queue: Jamie Madill <jmadill@chromium.org>
TBR=ynovikov@chromium.org,geofflang@chromium.org,jmadill@chromium.org
Change-Id: I6fdd34880a711ecc89d407ae90f3991584f65272
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: angleproject:2966
Reviewed-on: https://chromium-review.googlesource.com/c/1340199
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
d9ee8bf9
|
2018-11-12T20:04:20
|
|
Minor inlining optimization to Context draw calls.
Reduces draw call overhead by up to 3%.
Bug: angleproject:2966
Change-Id: Ie7ddb61b905fefe59a06a1528f0a3fde4accaf74
Reviewed-on: https://chromium-review.googlesource.com/c/1333608
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
0cc11c68
|
2018-10-12T18:07:18
|
|
StateManagerGL: Remove setGenericShaderState.
We can mutate the BitSetIterator as it clears dirty bits. This removes
the risk of doing a double state update. Improves the proformance of
the GL back-end state update.
Also do an early-out before calling syncDrawArraysState.
Bug: angleproject:2763
Change-Id: Idd25bdd67a6aceff05529a533260b661b07c2928
Reviewed-on: https://chromium-review.googlesource.com/c/1262740
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
ca8eda41
|
2018-10-18T18:41:56
|
|
Use dispatch table to optimize buffer binding.
Using a table of function pointers is faster than using a switch
followed by a function call. Also more aggressively inline binding
methods.
Based on contribution by mtavenrath@nvidia.com.
In total this patch sequence improves the performance of a buffer
binding perf test by up to 27%.
Test: BindingsBenchmark.Run/gl_100_objects_allocated_at_initialization
Bug: angleproject:2891
Change-Id: Iaab1e2a135b635bd72736d7d1d4271562c3a4ece
Reviewed-on: https://chromium-review.googlesource.com/c/1281783
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
|
|
956ab4d9
|
2018-10-10T16:13:03
|
|
Optimize several functions for the Program perf test.
This gives the same or slightly better performance in the ProgramDraw
perf test. Also only set the Program object as dirty when there are
dirty bits set in the Program itself.
Bug: angleproject:2877
Change-Id: I07b428b40d3e3c24e0a42c970524756b6dc3a30e
Reviewed-on: https://chromium-review.googlesource.com/c/1271475
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
6f755b21
|
2018-10-09T12:48:54
|
|
Use angle::Result in front-end. (Part 1)
This covers most of the hot paths used in draw calls. Gives in the
order of a 5% reduction in draw call overhead.
Bug: angleproject:2491
Change-Id: I2d53afb1163eaceed61fb9cd9ce6c1267c85c0fa
Reviewed-on: https://chromium-review.googlesource.com/c/1258149
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
785e8a0b
|
2018-10-04T17:42:00
|
|
Remove gl::LinkResult.
Instead of returning a small struct from LinkProgram calls we use
angle::Result. Linking can have 3 cases:
- the link was successful -> angle::Result::Continue
- the link failed -> angle::Result::Incomplete
- there was an internal error -> angle::Result::Stop
Note that any unexpected Incomplete is still an error. Each function
that accepts Incomplete must check explicitly.
This is the last user of ErrorOrResult.
Bug: angleproject:2491
Change-Id: Idba23be27efe4b561720a4bdd8fe486b40779497
Reviewed-on: https://chromium-review.googlesource.com/c/1255645
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@google.com>
|
|
14246813
|
2018-10-03T17:51:16
|
|
Update sampler directly on Texture change.
This saves a lot of extra work when updating a Texture.
Bug: angleproject:2763
Change-Id: I87e310ef4f847713123bd24711e1166949ff95d2
Reviewed-on: https://chromium-review.googlesource.com/c/1254043
Reviewed-by: Yuly Novikov <ynovikov@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
e3bb6b79
|
2018-10-03T17:51:15
|
|
Update Textures immediately on program change.
Instead of deferring this work, we can process it immediately. We have
access to the information we need in the prior and current states so
can optimize the update more effectively.
Bug: angleproject:2763
Change-Id: I77482b0ea490fddf403ccc4a4f32665301fd6e7a
Reviewed-on: https://chromium-review.googlesource.com/c/1254042
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
44a6fbfd
|
2018-10-02T13:38:56
|
|
Optimize resolveLink.
This changes the program query to resolve the link if required. If the
validation layer is skipped the link is resolved in the gl::Context.
Resolving the link on program query allows us to avoid resolving the
link on most of the gl::Program query APIs.
This improves inlining and particularly affects uniform update. It
fixes a performance regression introduced by the parallel shader
linking extension. Gives a 17% increased score on a uniform benchmark.
Also fixes two missing cases of checking for the extension in our
validation code.
Note that some bugs might still exist when the validation layer is
disabled.
Bug: angleproject:2851
Change-Id: I5d725eede3fa147cedf2ce0129791b3412a97a61
Reviewed-on: https://chromium-review.googlesource.com/1255509
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@google.com>
|
|
e25b8006
|
2018-09-20T13:39:49
|
|
Move sampler sync out of syncProgramTextures.
We only need to call the syncState for samplers when they are dirty.
Also includes changes to refactor out the sampler sync in GL. Adds
observer bindings so sampler sync is handled correctly in resource
sharing scenarios.
Bug: angleproject:2763
Change-Id: I762f0738ee7572ae29ce6bd5384a30aa9588c848
Reviewed-on: https://chromium-review.googlesource.com/1227797
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
99118c1c
|
2018-09-10T16:28:51
|
|
Improve dirtybits for compute
For D3D11 backend, dirtybits are used to sync and update state
for compute.
For GL backend, mProgramUniformBuffersDirty,
mProgramAtomicCounterBuffersDirty, mProgramImagesDirty are newly
applied.
Bug: angleproject:2814
Change-Id: Ia459ca26705b7c685179d64e0c934535bf16741f
Reviewed-on: https://chromium-review.googlesource.com/1215245
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>
|
|
0fdb956d
|
2018-09-17T17:18:43
|
|
Re-land "Inline and micro-optimize more for perf tests.""
Re-land fixes memory leaks.
Using a custom array instead of std::vector speeds up the resource
manager. One reason is because calls to size() are implemented in many
implementations as a difference between two pointers. This sub size
implementations are slower than storing a simple size variable in a
custom class.
Also includes more inlining of hot spots functions.
Also includes a small unit test class for ResourceMap. And an unrelated
but small test fix for TextureLimisTest. Also a small unrelated fix for
a Transform Feedback test.
Increase the scores of the draw call perf test with texture and buffer
bindings and the buffer binding perf test.
Bug: angleproject:2763
Change-Id: Ic2f0f689107b2bf05c63da2ed6bbc9f0feea63f7
Reviewed-on: https://chromium-review.googlesource.com/1229033
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
98a3e078
|
2018-09-17T19:40:04
|
|
Revert "Inline and micro-optimize more for perf tests."
This reverts commit 57ff6f95f143bd65a0c3d12d64773f274b9935f4.
Reason for revert: Memory leaks detected during roll in https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_asan_rel_ng/100284
Original change's description:
> Inline and micro-optimize more for perf tests.
>
> Using a custom array instead of std::vector speeds up the resource
> manager. One reason is because calls to size() are implemented in many
> implementations as a difference between two pointers. This sub size
> implementations are slower than storing a simple size variable in a
> custom class.
>
> Also includes more inlining of hot spots functions.
>
> Also includes a small unit test class for ResourceMap. And an unrelated
> but small test fix for TextureLimisTest. Also a small unrelated fix for
> a Transform Feedback test.
>
> Increase the scores of the draw call perf test with texture and buffer
> bindings and the buffer binding perf test.
>
> Bug: angleproject:2763
> Change-Id: I41c327987db27ac45e6a62579f01e1cdc22e396c
> Reviewed-on: https://chromium-review.googlesource.com/1171510
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Jamie Madill <jmadill@chromium.org>
> Commit-Queue: Jamie Madill <jmadill@chromium.org>
TBR=fjhenigman@chromium.org,ynovikov@chromium.org,jmadill@chromium.org,syoussefi@chromium.org
Change-Id: Ie047289c9bf23a842c3cbb9692c811da0534991c
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: angleproject:2763
Reviewed-on: https://chromium-review.googlesource.com/1228893
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
57ff6f95
|
2018-09-14T12:44:52
|
|
Inline and micro-optimize more for perf tests.
Using a custom array instead of std::vector speeds up the resource
manager. One reason is because calls to size() are implemented in many
implementations as a difference between two pointers. This sub size
implementations are slower than storing a simple size variable in a
custom class.
Also includes more inlining of hot spots functions.
Also includes a small unit test class for ResourceMap. And an unrelated
but small test fix for TextureLimisTest. Also a small unrelated fix for
a Transform Feedback test.
Increase the scores of the draw call perf test with texture and buffer
bindings and the buffer binding perf test.
Bug: angleproject:2763
Change-Id: I41c327987db27ac45e6a62579f01e1cdc22e396c
Reviewed-on: https://chromium-review.googlesource.com/1171510
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
bb2f2c43
|
2018-09-16T09:47:38
|
|
Squash State dirty bits for CHROMIUM_path_rendering
Currently we limit the maximum of State bits to 64 for performance
consideration. To accommodate new ES31 states, we have to pick and
squash some existing bits which are less hot.
Bug: angleproject:2814
Change-Id: I8fc19286bbf8d8273949da8d986a8b1587fa105e
Reviewed-on: https://chromium-review.googlesource.com/1227069
Reviewed-by: Jie A Chen <jie.a.chen@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>
|
|
63aa0e5b
|
2018-09-05T16:15:46
|
|
Fix 64-bit -> 32-bit implicit conversions in libangle.
../../third_party/angle/src/libANGLE/renderer/gl/StateManagerGL.cpp(910,63): warning: implicit conversion loses integer precision: 'size_t' (aka 'unsigned long long') to 'GLuint' (aka 'unsigned int') [-Wshorten-64-to-32]
const gl::ImageUnit &imageUnit = glState.getImageUnit(imageUnitIndex);
~~~~~~~~~~~~ ^~~~~~~~~~~~~~
../../third_party/angle/src/libANGLE/renderer/gl/StateManagerGL.cpp(914,30): warning: implicit conversion loses integer precision: 'size_t' (aka 'unsigned long long') to 'GLuint' (aka 'unsigned int') [-Wshorten-64-to-32]
bindImageTexture(imageUnitIndex, textureGL->getTextureID(), imageUnit.level,
~~~~~~~~~~~~~~~~ ^~~~~~~~~~~~~~
../../third_party/angle/src/libANGLE/renderer/gl/StateManagerGL.cpp(920,30): warning: implicit conversion loses integer precision: 'size_t' (aka 'unsigned long long') to 'GLuint' (aka 'unsigned int') [-Wshorten-64-to-32]
bindImageTexture(imageUnitIndex, 0, imageUnit.level, imageUnit.layered, imageUnit.layer,
~~~~~~~~~~~~~~~~ ^~~~~~~~~~~~~~
../../third_party/angle/src/tests/gl_tests/VertexAttributeTest.cpp(1080,66): warning: implicit conversion loses integer precision: 'GLsizeiptr' (aka 'long long') to 'GLuint' (aka 'unsigned int') [-Wshorten-64-to-32]
glVertexAttribFormat(mTestAttrib, 1, GL_FLOAT, GL_FALSE, inputRelativeOffset);
~~~~~~~~~~~~~~~~~~~~ ^~~~~~~~~~~~~~~~~~~
BUG=chromium:879657
Change-Id: Ic6e8e5ebc0dc5fd38c15a48a936ceafd5407bba8
Reviewed-on: https://chromium-review.googlesource.com/1208315
Commit-Queue: Will Harris <wfh@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
47f6dd0d
|
2018-08-10T13:36:32
|
|
Implement robust initialization for image texture
This change fixes a bug that the image texture initialization time is not
correct. It happens when we use compute shader to write data to an
uninitialized texture, then use readPixels to read the result which
results that texture initialization falls behind image store and
covers the real result.
Bug: angleproject:2766
Change-Id: I4e986972096857afc975c40dfa4d559a2f31194c
Reviewed-on: https://chromium-review.googlesource.com/1170569
Commit-Queue: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
7b2f0274
|
2018-08-27T10:16:08
|
|
Inline more buffer binding calls.
This splits checkObjectAllocation into the inline part and the
slow part. It uses ANGLE_INLINE to force the former to be inlined.
Also inlines and optimizes a few other buffer binding related checks.
Improves performance in a bindings performance test by about 20%.
Includes contributions from matavenrath@nvidia.com.
Test: Bindings_gl_100_objects_allocated_at_initialization
Bug: angleproject:2777
Change-Id: I71b07d72a8e81db7a90140ef84eca599a29239ed
Reviewed-on: https://chromium-review.googlesource.com/1190442
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
6d32cefd
|
2018-08-14T02:34:28
|
|
Use Observer pattern for Buffers in front-end.
This will allow us to react to state change notifications for validation
caching. It also cleans up some of the logic in the D3D11 State Manager.
Bug: angleproject:2747
Change-Id: I85ed6404206c2b9bf504d552cf5751be56e62146
Reviewed-on: https://chromium-review.googlesource.com/1172086
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
70aeda49
|
2018-08-20T12:17:40
|
|
Add gl::Program::syncState and dirty bits.
Currently this handles uniform block bindings. Cleans up some logic in D3D.
Bug: angleproject:2747
Change-Id: I8c2989738d50a77d6f6d90a9ff11dceab6d3129c
Reviewed-on: https://chromium-review.googlesource.com/1172085
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
bf5177d3
|
2018-08-21T12:58:20
|
|
Optimize element array binding.
This inlines some methods and code that was regressing binding
performance. Improves the score of a bindings benchmark by 15%.
Based on contributions by matavenrath@nvidia.com.
Test: bindings_gl_100_objects_allocated_at_initialization
Bug: angleproject:2777
Change-Id: Ied5fdbc67ced862a36c5145fff5ac94f93b40c21
Reviewed-on: https://chromium-review.googlesource.com/1181865
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
6c43a01a
|
2018-08-08T15:49:27
|
|
Make Context handle dirty texture events.
Moving from State to Context allows the Context to update the State
Cache class directly. It also calls through to the State class to
update the Texture cache. This consolidates notification events into
the Context class. This is also in line with how we handle state event
updates in other gl classes.
Bug: angleproject:2747
Bug: angleproject:2763
Change-Id: Iff7dc7e46ee8768819235ebd151707cd2a03dfc9
Reviewed-on: https://chromium-review.googlesource.com/1166143
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
4787d70b
|
2018-08-08T15:49:26
|
|
Use active textures mask in GL and Vulkan.
This also inlines a few accessors and changes the type of the texture
cache.
Bug: angleproject:2763
Change-Id: I82f3c508613f6284d835e8cb64808cfe26a56161
Reviewed-on: https://chromium-review.googlesource.com/1166142
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
7e4eff11
|
2018-08-08T15:49:26
|
|
Program: Add cache samplers and type masks.
This mask cleans up some of the iteration logic in
State::syncProgramTextures. It will make it easier to optimize this
function in the future. This will also make it easier to recompute
the sampler type validation.
Leads to a 5% improvement in State::syncProgramTextures.
Bug: angleproject:2747
Bug: angleproject:2763
Change-Id: Ic9a555df843ad23b4c562e6e4a2d425bee58a856
Reviewed-on: https://chromium-review.googlesource.com/1164306
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
cc73f241
|
2018-08-01T11:34:48
|
|
Micro-optimize some validation checks.
Prepares for caching hasMappedBuffer. Also inclines several checks for
faster speed.
Bug: angleproject:2746
Change-Id: I74f9408d7b41e245c3f58d367dd2cc8fbace4a7a
Reviewed-on: https://chromium-review.googlesource.com/1150762
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
a11819de
|
2018-07-30T10:26:01
|
|
Context: Use Observer pattern for container objects.
Container objects in this case are the Vertex Array and Read and Draw
Framebuffers. Instead of using a mutable dirty bit state with a const
function to notify of dirty bits we can use the Observer/Subject
pattern. This more cleanly allows us to do cache updates.
Bug: angleproject:1391
Change-Id: I88f863894ec3efa00322038f323a84850166107d
Reviewed-on: https://chromium-review.googlesource.com/1153399
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
9bf86f07
|
2018-07-26T11:46:34
|
|
Don't sync the read framebuffer on draw calls.
The read framebuffer may not be complete and be incapable of syncing.
Removed the generate syncDirtyObjects method so each caller must make sure they
are only syncing objects that are known to be valid for the operation.
BUG=angleproject:2737
Change-Id: Ia8edf3fca3a8369aa813be46ba99f6b50a36b2e6
Reviewed-on: https://chromium-review.googlesource.com/1151621
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
|
|
84c662b9
|
2018-07-12T15:56:55
|
|
Vulkan: Store program textures cache.
The program textures update needs to call into the incomplete textures
cache. Moving this step into an update mechanism allows us to avoid
some complexity of calling the incomplete textures cache inside the
ProgramVk class.
This enables further refactors for draw call shader patching and line
segment raster.
Bug: angleproject:2598
Change-Id: I0fd81ca58d2641572a2dd444992ccddd99ebc823
Reviewed-on: https://chromium-review.googlesource.com/1127301
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
82af620e
|
2018-06-22T10:59:52
|
|
ParallelCompile: Add entry points.
Add the extension text, entry points and validations.
BUG=chromium:849576
TEST=angle_end2end_tests
Change-Id: I4c06ee30e4f4fe9bb1c1fecada747b9c78fed0ea
Reviewed-on: https://chromium-review.googlesource.com/1103789
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>
|
|
4fb8a8bd
|
2018-06-01T16:47:57
|
|
Move the EGL-based initial state to the gl::State constructor.
This is a small refactor to allow State::initialize to be called without
needing the EGL context creation attributes.
BUG=angleproject:2464
Change-Id: Ifa167cc83f652435ecc00a0a73e4c6c4a3295430
Reviewed-on: https://chromium-review.googlesource.com/1083312
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
6b600645
|
2018-05-31T14:53:31
|
|
D3D11: Micro-optimize StateManager11::updateState.
This inlines several accessors.
Bug: angleproject:2574
Change-Id: I61d223dd2a8f08e5331ccefde02e6ce55f5a607d
Reviewed-on: https://chromium-review.googlesource.com/1067118
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
4166f014
|
2018-05-31T14:53:30
|
|
D3D11: Optimize Renderer11::drawArrays.
Uses inlining and more efficient computation re-use to streamline this
function.
Bug: angleproject:2575
Change-Id: Ib13e32811f56ec9a010ed66f298d4235e5c6807d
Reviewed-on: https://chromium-review.googlesource.com/1067120
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
ad3ae90a
|
2018-03-09T13:40:42
|
|
Use packed enums for QueryType.
BUG=angleproject:2169
Change-Id: I129a9d8e295859daa071a298dab9fe1895315cc0
Reviewed-on: https://chromium-review.googlesource.com/957318
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
|
|
7267aa65
|
2018-04-17T15:28:21
|
|
Optimize ValidateDrawAttribs: Part 3.
This is a small optimization for the WebGL compatibility mode.
Instead of scanning the list of attributes for a Transform feedback
conflict, it can quickly check a cached mask. This should save a lot
of cycles on the fast path.
Bug: angleproject:1391
Change-Id: Icb8d095493a629dbff0e93872357e5bf7c7458ae
Reviewed-on: https://chromium-review.googlesource.com/1011236
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
4bba74f0
|
2018-04-19T14:40:45
|
|
Refactor Texture::syncState to pass down the Context
Also returning a gl::Error everywhere.
Bug: angleproject:2478
Change-Id: Ic8cae0ee7aee318bb95b3588044c34c62707b578
Reviewed-on: https://chromium-review.googlesource.com/1020083
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>
|
|
abb09f12
|
2018-04-16T10:43:53
|
|
GLES1: Client vertex array pointer API
glVertexPointer
glNormalPointer
glColorPointer
glPointSizePointerOES
glTexCoordPointer
BUG=angleproject:2306
Change-Id: Id022b467ac998ea116130c5cec0c77afefb0dd4c
Reviewed-on: https://chromium-review.googlesource.com/1014381
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
0946393d
|
2018-04-04T05:26:59
|
|
Move Buffer Subject/Observer to front end.
This makes BufferImpl into an Observer Subject. It also refactors
the Vertex Array updates for the D3D11 backend use more of a dirty
bit coding style.
This change makes it so Buffer contents changes trigger front-end
dirty bits from the back-end, which may be undesirable.
Bug: angleproject:2389
Change-Id: Iac8ce1171284a86851c18cd1373ddf24fcefe40b
Reviewed-on: https://chromium-review.googlesource.com/979812
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
13b708f2
|
2018-03-21T12:14:10
|
|
GLES1: glAlphaFunc
BUG=angleproject:2306
Change-Id: I0bf229d3ab8a4a1217c12b434dcd8fa67d7cbadc
Reviewed-on: https://chromium-review.googlesource.com/973897
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
4d9d4839
|
2018-03-13T12:43:28
|
|
Optimize buffer deletion by returning early if buffer is not bound.
This improves angle_perftests
bindings_gl_100_objects_allocated_every_iteration by over 30%.
Bug: 820723
Change-Id: Icfaa0f54f7c61db6475948e3a95296f37ab4f7dc
Reviewed-on: https://chromium-review.googlesource.com/961302
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
dbb9c534
|
2018-02-25T19:58:24
|
|
GLES1: state: Define / initialize GLES1-specific states
Contains definitions of GLES1-specific states such as material /
lighting.
Tweaked the Color class for easier copying to uniforms / reading as a
float array.
This CL also adds the GLES1-specific state in GLES1State, which is then
part of the State class and is initialized to the spec's values if the
context major version is ES 1.
+ Some clang-format
BUG=angleproject:2306
Change-Id: I7fc3bd9a22ebf0ffcd98d931d0176f21e17b1c5c
Reviewed-on: https://chromium-review.googlesource.com/936424
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
e98b1b5d
|
2018-03-08T09:47:23
|
|
Framebuffer: Handle errors in checkStatus.
This pipes a lot more errors around in the Validation, where
they now will be caught.
Bug: angleproject:2372
Change-Id: Ibb4e47ddc932995a02dd92e10578b7a4097182a9
Reviewed-on: https://chromium-review.googlesource.com/954406
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
e8a93c6e
|
2018-01-04T18:02:24
|
|
New transform feedback buffer binding rules
Detects undefined behavior when a buffer is bound to a transform feedback
binding point and a non transform feedback binding point at the same time.
Also moves the transform feedback buffer generic binding point out of the
transform feedback object and into the context's global state, to match
driver behavior. This way binding a new transform feedback object does not
affect GL_TRANSFORM_FEEDBACK_BUFFER_BINDING which is similar to how VAOs
work with GL_ARRAY_BUFFER_BINDING.
Bug: 696345
Change-Id: If3b9306cde7cd2197a8ce35e10c3af9ee58da0b8
Reviewed-on: https://chromium-review.googlesource.com/853130
Commit-Queue: James Darpinian <jdarpinian@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
bc918e7d
|
2018-03-08T09:47:21
|
|
Return errors from gl::State::syncDirtyObjects.
This plumbs the front-end code, but doesn't implement Error handling
for any objects yet. Framebuffers and VertexArrays will get error
handling from their syncState methods.
Bug: angleproject:2372
Change-Id: I226049e25ac7af2bdd3c4787b67a5d297c0b9c4b
Reviewed-on: https://chromium-review.googlesource.com/954289
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
f0e89be6
|
2017-11-08T14:00:32
|
|
Use packed enums for the texture types and targets, part 1
In OpenGL there are two enum "sets" used by the API that are very
similar: texture types (or bind point) and texture targets. They only
differ in that texture types have GL_TEXTURE_CUBEMAP and target have
GL_TEXTURE_CUBEMAP_[POSITIVE|NEGATIVE]_[X|Y|Z].
This is a problem because in ANGLE we use GLenum to pass around both
types of data, making it difficult to know which of type and target a
variable is.
In addition these enums are placed somewhat randomly in the space of
OpenGL enums, making it slow to have a mapping from texture types to
some data. Such a mapping is in hot-code with gl::State::mTextures.
This commit stack makes the texture types and target enums be
translated to internal packed enums right at the OpenGL entry point
and used throughout ANGLE to have type safety and performance gains.
This is the first of two commit which does the refactor for all of the
validation and stops inside gl::Context. This was the best place to
split patches without having many conversions from packed enums to GL
enums.
BUG=angleproject:2169
Change-Id: Ib43da7e71c253bd9fe210fb0ec0de61bc286e6d3
Reviewed-on: https://chromium-review.googlesource.com/758835
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
6f683089
|
2018-02-28T00:35:16
|
|
Rename signal utils to Observer.
This completes the basic refactor to the Observer pattern.
Bug: angleproject:2372
Change-Id: I810deff7c7e39baa64b57ce2a79cd732b1af7c34
Reviewed-on: https://chromium-review.googlesource.com/940862
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
|
|
888081d5
|
2018-02-27T00:24:46
|
|
D3D11: Refactor dependent Framebuffer state changes.
Previously, when a state change would cause a Texture to
recreate its storage specific to D3D11, we would use a
dependent notification from RenderTarget11 to Framebuffer11
to re-check internal dirty bits. In this new method, we
instead set dirty bits on the gl::Frambuffer directly. This
also means we use fewer internal objects for these
notifications, because we share the same structures between
the D3D11 back-end notifications and the top-level notifications
we use for Robust init and Framebuffer completeness.
This also allows us to get rid of one "if" that we check on
every draw call in D3D11.
This also introduces a dirty bits guard concept - a shadow
set of dirty bits that is checked in dependent state changes
to ensure that extra bits aren't set inside syncState.
This also implements Framebuffer dirty bits for the D3D9
back-end. This has the side effect of cleaning up the
"null colorbuffer" D3D9 workaround.
Bug: angleproject:2372
Change-Id: Ie346d39030f4f6df583d735685b0babea4e745a8
Reviewed-on: https://chromium-review.googlesource.com/936691
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
d444255a
|
2018-02-27T22:03:47
|
|
Refactor signal utils into Observer pattern.
These types were over-generalized. All use cases featured
arrays of resources attached to single parent resources. The
channel ID is sufficient to identify the child resource in the
parent, and having variadic template arguments wasn't necessary.
Futhermore we can rename these types to use the common Observer
pattern. This should make them more readable to new developers.
Also update some classes to inherit from Subject instead of
having a member Subject. This cleans up the code in a few places.
This should lead to a simpler refactor to allow dependent dirty
bits notifications in the Vulkan back-end.
In the following patch the signal_utils files will be renamed. They
are not renamed in this patch to ensure git history is preserved.
Bug: angleproject:2372
Change-Id: I17a3f2c8d92afd4bb3cba2d378c3a2e8a6d7fb11
Reviewed-on: https://chromium-review.googlesource.com/936690
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
c405ae71
|
2017-12-06T14:15:03
|
|
Optimize Vertex Shader Attribute Type Validition
Improves ValidateVertexShaderAttributeTypeMatch by storing vertex
attributes types into masks for quick comparisons when needed. This shows
2% improvement to glDrawElements for the aquarium workload.
BUG=angleproject:2202
Change-Id: I87fa3d30c3d8cdba6dfd936cd1a41fd27b1c6b77
Reviewed-on: https://chromium-review.googlesource.com/814795
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
f414121d
|
2017-12-12T15:08:07
|
|
D3D11: Fix program uniform buffer dity bits.
There were actually three missing state synchronization points:
1. When re-binding a uniform buffer, we had no dirty bits.
2. When modifying the Program bindings, we also had no update.
3. When triggering a dependent state update via BufferData.
This fixes all the missing syncs and adds tests for the cases.
Bug: angleproject:1390
Change-Id: I2f05766545784248bcca8308ae34ac5720237ab1
Reviewed-on: https://chromium-review.googlesource.com/823150
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
a98a2811
|
2017-11-30T18:12:06
|
|
ES31: Add support for GL_DISPATCH_INDIRECT_BUFFER_BINDING binding point
BUG=angleproject:2270
TEST=dEQP-GLES31.functional.state_query.integer.dispatch_indirect_buffer_binding_*
Change-Id: I2c286c219c5c849d7ba417069f53f2d0fae6abf1
Reviewed-on: https://chromium-review.googlesource.com/799990
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
10a4d434
|
2017-11-28T14:46:26
|
|
ES31: Enable some dirty bits and dirty objects for compute pipeline
BUG=angleproject:2265
TEST=dEQP-GLES31.functional.shaders.builtin_var.compute.*
angle_end2end_test.ShaderStorageBufferTest31
.MultiStorageBuffersForMultiPrograms
Change-Id: Icc3df122602951a2328003c10a76696ab4c9f0d8
Reviewed-on: https://chromium-review.googlesource.com/792951
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
ded7923b
|
2017-11-28T15:21:11
|
|
GL backend: Only synchronize transform feedback state when it changes.
BUG=angleproject:2188
Change-Id: I5bfcc038c887dde0770564d103eb3cb234b248c9
Reviewed-on: https://chromium-review.googlesource.com/794100
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
1f679cc6
|
2017-11-29T18:06:00
|
|
Code refactoring for angle::BitSet and EXPECT_GL_TRUE/FALSE.
This change refactors two style issues to make it be consistent.
1) This CL uses "using" to replace "typedef" for all angle::BitSet<...>.
2) This CL uses EXPECT_GL_TRUE/FALSE to replace EXPECT_EQ for bool comparison.
BUG=angleproject:2005
Change-Id: I4afad92313ea2457bbfedf80f917a5873d7f29ee
Reviewed-on: https://chromium-review.googlesource.com/795871
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
e51ba63b
|
2017-11-21T11:45:25
|
|
Optimize Texture initialization tracking with extra caching.
In both Texture and State, track when all images/textures are
initailized so that State::clearUnclearedActiveTextures can early-exit.
Improves performance on the WebGL Aquarium by 26% (23->29 FPS) with 30000
fish.
BUG=angleproject:2188
Change-Id: Ie2860a81d7be19ee87262325d8cf27bde43f80b8
Reviewed-on: https://chromium-review.googlesource.com/782339
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
acf2f3ad
|
2017-11-21T19:22:44
|
|
Apply Chromium style fixes.
This addresses several minor code quality issues that are validated
in Chromium, but not yet applied to ANGLE:
* constructors and destructors must be defined out-of-line
* auto is not allowed for simple pointer types
* use override everywhere instead of virtual
* virtual functions must also be defined out-of-line
Slightly reduces binary size for me (~2k on Win, 150k on Linux).
Bug: angleproject:1569
Change-Id: I073ca3365188caf5f29fb28d9eb207903c1843e6
Reviewed-on: https://chromium-review.googlesource.com/779959
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
cac94a91
|
2017-11-10T10:09:32
|
|
Optimize ShaderVariable::isBuiltIn.
This makes the check a bit faster, by inlining the prefix check.
Also some cleanups to ValidateVertexShaderAttributeTypeMatch.
BUG=angleproject:2202
Change-Id: Ifeab4cd85a91a1639a461f44776a68ac98c5bd79
Reviewed-on: https://chromium-review.googlesource.com/761240
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
336129f6
|
2017-10-17T15:55:40
|
|
Use a packed enum for buffer targets.
BUG=angleproject:2169
Change-Id: I4e08973d0e16404b7b8ee2f119e29ac502e28669
Reviewed-on: https://chromium-review.googlesource.com/723865
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
cda6af19
|
2017-10-30T19:20:37
|
|
Split pixelBuffer from pack/unpack state
This will refactor will help use packed enums for buffer targets.
BUG=angleproject:2169
Change-Id: Ie7ed3e105f89457c67027e6598d7e29503ad355c
Reviewed-on: https://chromium-review.googlesource.com/745181
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
29a20992
|
2017-11-06T18:23:16
|
|
Add back dirty bits for pack / unpack buffer binding
BUG=angleproject:2169
Change-Id: I8cee9e1c7565d90022216d62b45f0eec5bc51f14
Reviewed-on: https://chromium-review.googlesource.com/755461
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
|
|
c67323a9
|
2017-11-02T23:11:41
|
|
Squash State dirty bits to below 64.
This will dramatically speed up performance on 64-bit systems due to
the efficiency of angle::BitSet64. Improves performance of the GL
back-end on benchmarks quite a bit.
Squashes the Pack and Unpack states together. Also moves the current
value dirty bits to a single dirty bit, with a separate set for each
attribute. Also squashes a multisample dirty bit.
Also fixes an incorrect dirty bit in StateManagerGL.
We may want to implement a semi-fast version of BitSet64 for 32-bit
systems if we find performance is not satisfactory.
BUG=angleproject:2188
Change-Id: I4616782bf75413252ede3e3ac752b9ccdb9aab49
Reviewed-on: https://chromium-review.googlesource.com/722423
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
a59fc19f
|
2017-11-02T12:57:58
|
|
Use active textures mask with robust init.
This should speed up clearUnclearedActiveTextures considerably.
It was showing up as a hotspot when running the aquarium demo with
the passthrough command decoder.
Also rename the complete textures mask in gl::State to an active
textures mask, since it includes incomplete textures.
BUG=angleproject:2188
Change-Id: Idf020fc49c1e74f17a8005c3b88516829767b84c
Reviewed-on: https://chromium-review.googlesource.com/722421
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
2e568cfb
|
2017-09-18T17:05:22
|
|
Add generator for packed GL enums.
For testing this also converts two unimportant GLenums, gl::BufferUsage
and gl::CullModeFace.
BUG=angleproject:2169
Change-Id: If1e86a97d0fed3fd567303aca6506ec579503076
Reviewed-on: https://chromium-review.googlesource.com/688000
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
42975644
|
2017-10-12T12:31:51
|
|
Move incomplete texture logic to shared helper.
The incomplete texture handling is similar between the D3D and Vulkan
back-ends. We create 1x1 textures, initialize them to black, and bind
them when we detect incomplete textures. We would also bind incomplete
textures when we detect feedback loops. In the GL back-end, we
wouldn't detect feedback loops, and would allow the driver to handle
incompleteness.
Instead implement this in a shared helper class, and do the feedback
loop detection in the front-end for every back-end. This makes our
behaviour more consistent between back-ends, and prevents undefined
behaviour.
Because initializing multisample textures is tricky (they
can't be updated with TexImage calls) we do a bit of a workaround so
the back-end can clear the incomplete multisample texture initially.
This progresses the initial Vulkan textures implementation.
BUG=angleproject:2167
Change-Id: I79ddcc0711fcc986f2578a52ac6f701231d241ac
Reviewed-on: https://chromium-review.googlesource.com/700993
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
05b35b21
|
2017-10-03T09:01:44
|
|
D3D11: Lazy robust resource init.
This patch moves the robust resource init logic to the GL front-end.
Instead of initializing texture resources immediately on creation in
D3D11, it defers the clear until before a draw call in some cases, or
skips the update if we can determine if a texture (or other resource)
has been fully initialized.
Currently lazy init is only implemented for Textures, Renderbuffers,
and Surfaces.
Various places where lazy resource init is triggered:
* Framebuffer operations (Draw, Blit, CopyTexImage, Clear, ReadPixels)
* Texture operations (SubImage, GenerateMipmap, CopyTexImage)
Some efficiency gains remain to be implemented, such as when a
SubImage call fills the entire object. Similarly for Blit, and a few
other operations. In these cases we can skip lazy init as an
optimization. Edge cases with EGLImage are mostly untested.
BUG=angleproject:2107
Change-Id: I2bf3a69b1eae0d4feeb5b17daca23451f1037be8
Reviewed-on: https://chromium-review.googlesource.com/576058
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
db342276
|
2017-09-27T10:21:45
|
|
ES31: Implement glSampleMaski on OpenGL
This patch intends to implement glSampleMaski on OpenGL
backends.
Refers to: https://chromium-review.googlesource.com/c/487603
BUG=angleproject:1592
TEST=dEQP-GLES31.functional.state_query.*.sample_mask*
TEST=dEQP-GLES31.functional.texture.multisample.*.sample_mask*
Change-Id: If5ddd6ab57259593919a482be80fbdbe29f6f54d
Reviewed-on: https://chromium-review.googlesource.com/573727
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
0f80ed86
|
2017-09-19T00:24:56
|
|
Improve speed of iterating dirty textures.
We had a performance regression in the Textures benchmark. What the
test was doing was iterating over all possible texture state,
ensuring the active texture was dirty every frame. This is an attempt
to improve on the speed by not doing as much resetting work in
State::syncProgramTextures. It introduces an active textures mask to
speed iteration over the active texture set.
Also makes a refactoring change to Context to make it easier to limit
caps to an implementation maxium. The number of active textures is
limited to 64 so they easily fit in the bitset mask, with a limit of
32 per shader stage. No mask is currenly kept for compute shaders.
With the fix the performance should be about the same as before (which
is good, as the test always sets the textures dirty).
Test: TexturesBenchmark.Run/gl_8_textures_5_rebind_3_state_8_mips
BUG=chromium:765363
BUG=angleproject:1387
Change-Id: I8bcf95be3671195373573f89f406edaba40aa1be
Reviewed-on: https://chromium-review.googlesource.com/670279
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
a336b90f
|
2017-08-02T16:05:21
|
|
ES31: Impl program pipeline object management entries for GL backend.
The program pipeline object management entries are:
GenProgramPipelines
DeleteProgramPipelines
BindProgramPipeline
IsProgramPipeline
BUG:angleproject:2123
Change-Id: I114d054b90caf2ee3f9befef7439552a1c309bc4
Reviewed-on: https://chromium-review.googlesource.com/629978
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
d68248be
|
2017-09-11T14:34:14
|
|
Update Texture dirty bits documentation.
BUG=angleproject:1387
Change-Id: Ie2277874acb9e7a3eed4bce327f2c08750213967
Reviewed-on: https://chromium-review.googlesource.com/660419
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
81c2e253
|
2017-09-09T23:32:46
|
|
Add top-level dirty bits for texture and samplers.
These will have to be fleshed out in the back-ends.
Also currently uses a single bit for all the bindings, and we can
extend this to more fine-grained updates in the future.
This patch implements top-level updates for texture completeness.
Sampler completeness caches are removed from the Texture class, and
replaced by a cache in the gl::State class. The State class also
keeps a channel binding to the bound textures so it can be notified
when textures might change from complete <-> incomplete.
In future CLs we skip updating back-ends if texture state doesn't
change.
BUG=angleproject:1387
Change-Id: If580b4851303c86f3240e62891f5f6047eefb6a2
Reviewed-on: https://chromium-review.googlesource.com/648053
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
a779b610
|
2017-07-24T11:46:05
|
|
Smart caching of VAO input layout.
Don't recompute the cached attribute layout for a program if the
vertex array info hasn't changed. We can use the Serial class to
know when a vertex array has identical state.
BUG=angleproject:1156
Change-Id: Ia11f6ac268f63c3299f6d6d80c2866009cb8429c
Reviewed-on: https://chromium-review.googlesource.com/529768
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
dde78e8c
|
2017-05-22T14:13:27
|
|
ES31: Implement Vertex Attrib Binding on OpenGL
This patch intends to implement Vertex Attrib Binding on OpenGL
back-ends:
1. Add supports for updating vertex attributes by Vertex Attrib
Binding APIs.
2. Refactor the process of updating vertex attribtues in class
VertexArray to make it easier to implement this feature.
BUG=angleproject:1593
TEST=dEQP-GLES31.functional.vertex_attribute_binding.*
Change-Id: I800e61518c552b94b84c415895ad31668b0a84b2
Reviewed-on: https://chromium-review.googlesource.com/510251
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
c43be720
|
2017-07-13T16:22:14
|
|
Implement ANGLE_program_cache_control extensions.
This will give the browsers the ability to control the cache size,
query and populate the contents, and trim cache contents on memory
pressure.
BUG=angleproject:1897
Change-Id: I6edaa7d307b890223db98792d5b074e4a7fdfaa4
Reviewed-on: https://chromium-review.googlesource.com/563606
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
|
|
65ec0b2e
|
2017-03-28T16:10:52
|
|
ES31: Add support for bindImageTexture on GL backend
This patch refers to https://chromium-review.googlesource.com/c/380636/
BUG=angleproject:1987
Change-Id: If621eed6ecaa7298214843a2a133801ca1487b03
Reviewed-on: https://chromium-review.googlesource.com/462088
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
|
|
4928b7ca
|
2017-06-20T12:57:39
|
|
Proliferate gl::Context everywhere.
This gives the D3D back-end access to the GL state almost anywhere.
This uses the onDestroy hook for Textures to push errors up from
destructors, although they still don't quite make it to the Context.
There are places, such as in EGL object (Context/Surface) destruction,
where we end up calling through to GL implementation internals without
having access to a gl::Context. We handle this via a proxy Context
to a Display, basically a null context, that has access to impl-side
state like the Renderer pointer if necessary. It does not have access
to the normal GL state.
Also Pass gl::Context to RefCountObject::release(). Since we're using
destroy() methods now, we should not ever call the destructor directly.
BUG=angleproject:1156
Change-Id: Ie4c32ad6bf6caaff0289901f30b5c6bafa2ce259
Reviewed-on: https://chromium-review.googlesource.com/529707
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
231c7f56
|
2017-04-26T13:45:37
|
|
Apply clang-format to many files.
This cleans up the formatting in many places.
BUG=None
Change-Id: I6c6652ebc042f1f0ffecced53582d09d66b4f384
Reviewed-on: https://chromium-review.googlesource.com/487884
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
876429b7
|
2017-04-20T15:46:24
|
|
Update gl2.h and update entry points.
Some method signatures were updated. Types like GLclampf and GLvoid
were replaced with other equivalents.
BUG=angleproject:1309
Change-Id: I05e8e2072c5a063d87ad96a855b907424661e680
Reviewed-on: https://chromium-review.googlesource.com/475011
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|