|
f2315dbe
|
2024-10-23T13:34:39
|
|
Reland: Vulkan: Update checks for promoted extensions
This relands commit c2219ef9ec03277b6f8ad18d6d1ade41ddf88ba6.
Features that depend on promoted extensions are now enabled when
the extensions are present, OR the version of Vulkan that they were
promoted to is present. This commit checks for Vulkan 1.1 because
that is the version currently supported.
In the reland, the promoted extensions are not added to the list of
extensions to be enabled, because the underlying driver may not have
exposed them as extensions at all.
Bug: angleproject:374483183
Change-Id: I98e004d3fb4ed673984859ab63d40702f877b535
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5955635
Reviewed-by: Jason Macnak <natsu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
2dcc80dd
|
2024-10-17T13:59:23
|
|
Vulkan: allocateDescriptorSet to avoid repeated try on same pool
DynamicDescriptorPool::allocateDescriptorSet has a few steps. It first
tries to allocate from the same pool. Then tries to allocate from
mCurrentPoolIndex and then loops all existing pools. This CL keeps the
same basic logic, but avoids repeated tries on the same pool.
Bug: angleproject:372268711
Change-Id: Ic3099ac8c68688fe9afe452f808be29ac9063d51
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5926182
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
a05fc2bc
|
2024-10-23T13:36:54
|
|
Revert "Vulkan: Update checks for promoted extensions"
This reverts commit 9c1f96b8606a564b565d37208268b9b06efff03e.
Reason for revert: Still fails on Cuttlefish (which uses SwiftShader)
Bug: angleproject:374975259
Original change's description:
> Vulkan: Update checks for promoted extensions
>
> Features that depend on promoted extensions are now enabled when
> the extensions are present, OR the version of Vulkan that they were
> promoted to is present. This commit checks for Vulkan 1.1 because
> that is the version currently supported.
>
> Bug: angleproject:374483183
> Change-Id: I17d7956ac8a604f531e020e77c6f532e616b9dd5
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5954316
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Bug: angleproject:374483183
Change-Id: I1bef03f99b044bca91950a62e4015630a7c91530
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5956483
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Auto-Submit: Cody Northrop <cnorthrop@google.com>
|
|
|
9a4c7495
|
2024-10-15T13:05:28
|
|
Vulkan: Add feature flag to enable descriptorSet cache
So that we can disable it to compare the performance difference.
Bug: angleproject:372268711
Change-Id: I02da254e5d58815741080634a2dd005617aa7432
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5936135
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
31c80bbf
|
2024-10-17T10:56:16
|
|
Vulkan: Avoid redundant work in updateFullActiveTextures
ContextVk keeps mActiveTexturesDesc, which gets updated by
UpdatePreCacheActiveTextures(). This is only used for cache lookup. When
there is a cache miss, we end up call updateFullActiveTextures() which
recomputes DescriptorSetDesc again, which is redundant work.
This CL removes mActiveTexturesDesc from ContextVk.
UpdatePreCacheActiveTextures has been changed to be a
DescriptorSetDescBuilder method so that it can directly update the
mDesc. updateFullActiveTextures has been renamed to
updateActiveTexturesForCacheMiss which avoid mDesc calculation.
updateFullActiveTextures is still kept for now which will be used in
next CL when cache is disabled.
Bug: b/372268711
Change-Id: Ic9a0cdaa7cefca5f72b599d26d079cef14888f07
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5905766
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
dd54eeec
|
2024-10-11T13:26:46
|
|
Reland "Vulkan: Track GPU progress for individual DescriptorSet"
This is a reland of commit 292102944add2ab30f4aa12a971cac456cc7726b
with the fix of garbage being added back to garbage list.
Original change's description:
> Vulkan: Track GPU progress for individual DescriptorSet
>
> Right now ProgramExecutableVk keeps VkDescriptorSet object, and
> DescriptorSetHelper is created when a cache entry becomes invalid.
> Further, DescriptorSetCache keeps the cache of {VkDescriptorSet,
> RefCountedDescriptorPoolHelper} pair. So we are having three different
> type of objects at different stages of life: VkDescriptorSet,
> DescriptorSetHelper, and {VkDescriptorSet,
> RefCountedDescriptorPoolHelper. This CL makes DescriptorSetHelper at
> creation and at cache and at garbage. With this change, you have a
> reference counted DescriptorSetHelper object (i.e, DescriptorSetPointer)
> during entire life cycle and is passed around between cache and program
> as is. This CL is preparation for the future CL where we may disable
> cache for descriptorSet. The descriptorSet will be added to garbage list
> and reused constantly without go through the cache code. We need to
> track the individual descriptorSet with ResourceUse so that it won't
> reuse until GPU is finished. This CL is making DescriptorSetHelper a GPU
> tracking object so that it will still just work when cache is disabled.
>
> Bug: angleproject:372268711
> Change-Id: I1cfb77cc5069b202d870388fd8809e265cdca90b
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5918586
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Commit-Queue: Charlie Lao <cclao@google.com>
> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Bug: angleproject:372268711
Change-Id: Ic920f99cc78cde1e94690bdbee3b885844fa155b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5954701
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
60da450e
|
2024-09-20T17:56:55
|
|
CL: Implicit cmd queue submit on release
From OpenCL spec (3.0):
https://registry.khronos.org/OpenCL/specs/3.0-unified/html/OpenCL_API.html#_flush_and_finish
clReleaseCommandQueue perform an
implicit flush of the command-queue.
Bug: angleproject:8639
Change-Id: I0a3b0f3c602075fd2a6e1ac0310345b3d568df64
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5410931
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Austin Annestrand <a.annestrand@samsung.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
45cc47af
|
2024-10-22T21:41:22
|
|
Revert "Vulkan: Track GPU progress for individual DescriptorSet"
This reverts commit 292102944add2ab30f4aa12a971cac456cc7726b.
Reason for revert: Causing bot failure in later CLs
Original change's description:
> Vulkan: Track GPU progress for individual DescriptorSet
>
> Right now ProgramExecutableVk keeps VkDescriptorSet object, and
> DescriptorSetHelper is created when a cache entry becomes invalid.
> Further, DescriptorSetCache keeps the cache of {VkDescriptorSet,
> RefCountedDescriptorPoolHelper} pair. So we are having three different
> type of objects at different stages of life: VkDescriptorSet,
> DescriptorSetHelper, and {VkDescriptorSet,
> RefCountedDescriptorPoolHelper. This CL makes DescriptorSetHelper at
> creation and at cache and at garbage. With this change, you have a
> reference counted DescriptorSetHelper object (i.e, DescriptorSetPointer)
> during entire life cycle and is passed around between cache and program
> as is. This CL is preparation for the future CL where we may disable
> cache for descriptorSet. The descriptorSet will be added to garbage list
> and reused constantly without go through the cache code. We need to
> track the individual descriptorSet with ResourceUse so that it won't
> reuse until GPU is finished. This CL is making DescriptorSetHelper a GPU
> tracking object so that it will still just work when cache is disabled.
>
> Bug: angleproject:372268711
> Change-Id: I1cfb77cc5069b202d870388fd8809e265cdca90b
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5918586
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Commit-Queue: Charlie Lao <cclao@google.com>
> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Bug: angleproject:372268711
Change-Id: I4d3c34058d100112a098144276b52c0faf8d593a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5955529
Auto-Submit: Charlie Lao <cclao@google.com>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
|
|
9c1f96b8
|
2024-10-22T12:58:27
|
|
Vulkan: Update checks for promoted extensions
Features that depend on promoted extensions are now enabled when
the extensions are present, OR the version of Vulkan that they were
promoted to is present. This commit checks for Vulkan 1.1 because
that is the version currently supported.
Bug: angleproject:374483183
Change-Id: I17d7956ac8a604f531e020e77c6f532e616b9dd5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5954316
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
99ba07d3
|
2024-03-26T00:58:57
|
|
CL/Vulkan: Implement createImage
Enabling functionality of:
- clCreateImage
Tests-Passing: OCLCTS.test_api get_image<1d|2d|3d>_info
Bug: angleproject:42266936
Signed-off-by: Rafay Khurram <r.khurram@samsung.com>
Change-Id: I0281f092bff13cdd81b87d596fdd15b33dda7e46
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5796527
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
29210294
|
2024-10-11T13:26:46
|
|
Vulkan: Track GPU progress for individual DescriptorSet
Right now ProgramExecutableVk keeps VkDescriptorSet object, and
DescriptorSetHelper is created when a cache entry becomes invalid.
Further, DescriptorSetCache keeps the cache of {VkDescriptorSet,
RefCountedDescriptorPoolHelper} pair. So we are having three different
type of objects at different stages of life: VkDescriptorSet,
DescriptorSetHelper, and {VkDescriptorSet,
RefCountedDescriptorPoolHelper. This CL makes DescriptorSetHelper at
creation and at cache and at garbage. With this change, you have a
reference counted DescriptorSetHelper object (i.e, DescriptorSetPointer)
during entire life cycle and is passed around between cache and program
as is. This CL is preparation for the future CL where we may disable
cache for descriptorSet. The descriptorSet will be added to garbage list
and reused constantly without go through the cache code. We need to
track the individual descriptorSet with ResourceUse so that it won't
reuse until GPU is finished. This CL is making DescriptorSetHelper a GPU
tracking object so that it will still just work when cache is disabled.
Bug: angleproject:372268711
Change-Id: I1cfb77cc5069b202d870388fd8809e265cdca90b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5918586
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
47c66901
|
2024-10-21T12:47:22
|
|
Vulkan: Set gl_Layer to 0 if the framebuffer is not layered
Bug: angleproject:372390039
Change-Id: I29067c9488e06f6dd2e90f207fecb843267fb77c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5949263
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
e869b426
|
2024-10-22T10:04:06
|
|
Revert "Removed checks for promoted extensions"
This reverts commit c2219ef9ec03277b6f8ad18d6d1ade41ddf88ba6.
Reason for revert: Possible cause of boot failures on cuttlefish
Bug: angleproject:374975259
Original change's description:
> Removed checks for promoted extensions
>
> Removed checks for Vulkan extensions that have been promoted into Vulkan
> 1.1, which is already the minimum required version for ANGLE. The
> dependent features now work when the extensions are missing because the
> extensions are no longer necessary.
>
> Bug: angleproject:374483183
> Change-Id: Iad92b33a8e724eaa8f4bfcbe406f5b6b8536d6d8
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5944678
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Charlie Lao <cclao@google.com>
> Commit-Queue: Charlie Lao <cclao@google.com>
Bug: angleproject:374483183
Change-Id: I92eb6681c6d1b8f7e2d0832bb4e1cbf4597bd1c6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5952279
Commit-Queue: Cody Northrop <cnorthrop@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
4bdcdf0d
|
2024-10-16T11:44:49
|
|
Vulkan: Switch RefCountedDescriptorPoolBinding to use SharedPtr
This mostly a clean up. RefCountedDescriptorPoolBinding is replaced with
DescriptorPoolPointer, which is defined as
SharedPtr<DescriptorPoolHelper>. It has more intuitive semantics to use.
Bug: angleproject:372268711
Change-Id: I0397111b5228e896c1d226e00930851319d955a0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5938947
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
c52e8493
|
2024-10-11T14:21:19
|
|
Vulkan: Switch DescriptorPoolPointer to use SharedPtr
To make the reference counting easier to maintain, SharedPtr class is
added to automatically tracking the reference counting and object
creation/destruction. Right now we have BindingPointer class doing
similar things except it does not automatically manage the object
create/destroy, which makes it less robust as well as redundant code to
manage object life cycle. The other problem with BindingPointer is that
it does not work with normal assign/copy operator which made it hard to
read and use. SharedPtr uses exact same API semantics as
std::shared_ptr, which makes the reference counting very easy to use.
The main difference of SharedPtr and std::shared_ptr is that it does not
use any atomic or lock since it assumes user only uses it under thread
safe environment, which ANGLE's backend is.
This CL also changes mDescriptorPools to mDynamicDescriptorPools to make
it clear that it is dynamic descriptor pool not the descriptor pool.
This is also preparation CL for the next CL where we will use SharedPtr
to manage DescriptorSetHelper life cycle, which otherwise a bit
complicated to manually manage.
Bug: angleproject:372268711
Change-Id: I1033d9bf259bbc075a9b374d8a28e1f67d889873
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5926183
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
a19f0947
|
2024-10-17T22:42:30
|
|
Vulkan: Cache depth- and stencil-only views
Existing depth/stencil blit and resolve paths created temporary depth-
and stencil-only views. For
GL_ARM_shader_framebuffer_fetch_depth_stencil, such views are needed as
well.
In preparation for that extension, this change adds depth- and
stencil-only views to ImageViewHelper and allows them to be retrieved
through RenderTargetVk. The blit and resolve paths are consequently
simplfied as a side-effect.
Bug: angleproject:352364582
Change-Id: Ia822efb44ca7c82f63afce904eb19dd1bed02ff5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5938149
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
c2219ef9
|
2024-10-20T04:49:44
|
|
Removed checks for promoted extensions
Removed checks for Vulkan extensions that have been promoted into Vulkan
1.1, which is already the minimum required version for ANGLE. The
dependent features now work when the extensions are missing because the
extensions are no longer necessary.
Bug: angleproject:374483183
Change-Id: Iad92b33a8e724eaa8f4bfcbe406f5b6b8536d6d8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5944678
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
323187d9
|
2024-10-11T13:48:51
|
|
Vulkan: Fix color attachment limit with framebuffer fetch
ANGLE incorreclty assumed that the input descriptor limit is at least as
big as the color attachment limit. This is not true on Intel/windows
where 8 color attachments are available but only 7 input descriptors.
With this change, the color attachment limit is dropped to 7 in such a
case so that framebuffer fetch can continue to be supported.
Bug: angleproject:372873263
Change-Id: If836563b47399a23b293b74815f6bccb21aaf47c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5919759
Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
3fa74223
|
2024-03-26T17:13:52
|
|
Vulkan: Add check for VK_EXT_external_memory_host extension
Tests-Passing: GN Build
Bug: angleproject:42266936
Change-Id: Ie30f1cee939bc8a4b98d37f718e4526156916f05
Signed-off-by: Rafay Khurram <r.khurram@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5796524
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
897a5654
|
2024-10-16T10:32:16
|
|
CL: Rename isSet/isNotSet to intersects/excludes
In particular, `isSet` was checking if _any_ bit was set instead of if
_all_ bits were set. This was a point of confusion.
Bug: angleproject:42266936
Change-Id: I76211a423f304fb4641ceb9cbfbb0bea4ff33ccf
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5937411
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
e40d8581
|
2024-10-16T10:57:39
|
|
Vulkan: Fix render pass revival vs framebuffer fetch and DR
Bug: angleproject:352364582
Change-Id: I86548251fc1dec75031a23e3461bf296c852919c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5937412
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
|
7beb008d
|
2024-10-16T10:04:43
|
|
Vulkan: Disable dynamic rendering on Nvidia
Bug: angleproject:372883691
Change-Id: I71b45ea67a725a5a55d69b17329866000a1f37e3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5938740
Reviewed-by: Alexey Knyazev <lexa.knyazev@gmail.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
0dbe85f3
|
2024-10-15T13:54:36
|
|
Increase the size of vector WriteImages to max
ANGLE limits the size of vector which represents the write images when
resolving images. So when blit a multisample buffer to mrt, the sum of
write images is more than 1 and app will abort while checking the size
of the vector.
This patch increases the size of vector WriteImages to max.
Add end2end test to test blit multisampled framebuffer to MRT
framebuffer.
Bug: angleproject:361369302
Change-Id: I2d892bcd3411f2bca2ff514f6f0b6055d872668a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5872512
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
91391c06
|
2024-10-02T16:07:02
|
|
Vulkan: Vertex attribute hole crash workaround
* Added condition to bypass reading from a streaming attribute
if the source pointer is null.
* Added unit test that crash if a vertex pointer is not defined
for an enabled vertex attribute.
* VertexAttribPointerCopyBufferFromInvalidAddress
* Credit for the original test: tingwei.guo
Bug: angleproject:359729255
Change-Id: I2592fed66f0eba8c7003ec02cc8ca802833f23b3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5899978
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
a1584f49
|
2024-10-11T21:17:32
|
|
Vulkan: Qualify framebuffer fetch with "Color"
In preparation for depth/stencil framebuffer fetch, many framebuffer
fetch symbols are affixed with Color to indicate that they pertain to
color framebuffer fetch logic.
Bug: angleproject:352364582
Change-Id: I86000ada5e6ef47387dec0b6a3fca589d816cdc2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5926593
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
1608d0be
|
2024-10-10T16:53:15
|
|
Vulkan: Isolate framebuffer fetch no-RP-break optim from DR
Prior to [1], changes to framebuffer fetch usage by shaders caused a
render pass break. This was due to a limitation of render pass
compatibility rules. It also caused other headache, such as needing to
clear the render pass cache, recreating pipelines etc.
[1]:https://chromium-review.googlesource.com/c/angle/angle/+/3697308
In [1] an important optimization was implemented for tiling GPUs where
ANGLE permanently switched to framebuffer fetch mode on first
encountering framebuffer fetch use. From that point on, ANGLE would
always make every render pass framebuffer fetch compatible.
In reality, the render pass break was unnecessary, which became apparent
with dynamic rendering (for example that whether the render pass
includes input attachments has no bearing on a pipeline that doesn't use
input attachments at all). In [2], dynamic rendering kept the render
pass break + permanent switch behavior for simplicity.
[2]:https://chromium-review.googlesource.com/c/angle/angle/+/5637155
This change untangles the optimization done for legacy render passes
from dynamic rendering, allowing dynamic rendering to start every render
pass without framebuffer fetch and enable it later if a framebuffer
fetch program is used.
This is in preparation for supporting depth/stencil framebuffer fetch,
where a perma-switch is troublesome (for example in combination with
read-only depth/stencil feedback loops).
Bug: angleproject:352364582
Change-Id: I31221cf22a28d58b9b2bf188e9c0b786cd0fe3d2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5923120
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
|
|
8f367854
|
2024-10-13T13:45:17
|
|
Vulkan: Refactor ImageCopy shader
- Refactor common shader code into "ImageCopy.inc"
- Add ImageCopyFloat shader that supports only float src and dst formats
- Perform YUV and multisampled image copies with ImageCopyFloat
Bug: angleproject:372059358
Change-Id: I34bbaf49f98920494d9ed9e1fd290b413a89ea13
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5927276
Commit-Queue: mohan maiya <m.maiya@samsung.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
78a694a1
|
2024-10-10T14:08:31
|
|
Bugfix for ms_to_ss in dynamic rendering
VkMultisampledRenderToSingleSampledInfoEXT should be added into
VkRenderingInfo.pNext when using dynamic rendering pass.
Bug: angleproject:370153437
Change-Id: If9377362d95c60e610ec978eb684918de9807342
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5922915
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
68de0042
|
2024-10-08T05:41:37
|
|
Vulkan: Support glCopyTexImage2D from MSAA default framebuffer
Add support for glCopyTexImage2D when the source is an MSAA
default framebuffer by enhancing UtilsVk::copyImage(...) to
support multisampled source images.
Bug: angleproject:372059358
Tests: MultisampleTestES3.CopyTexImage2DFromMsaaDefaultFbo*
Change-Id: I9b5d39f6f75ca8056f8be18b1e68eb0161c99903
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5914629
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
|
|
|
b724eb0e
|
2024-10-10T16:51:11
|
|
Vulkan: Fix assert with overlay and not dynamic rendering
The render pass started by UtilsVk for the overlay was not marked as
belonging to the default framebuffer, causing an incorrect assertion
failure where the final image layout was to be set to PRESENT_SRC.
Bug: angleproject:42267038
Change-Id: I6e7a9357378d6f9a6e907e0ba25fb64dc8f7540e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5923119
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
62f33a5c
|
2024-10-09T16:06:54
|
|
Vulkan: Make retainResource for descriptorSetPool consistent
Right now the cache hit and cache miss case are handled differently.
This CL makes it consistent. Now Caller of
DynamicDescriptorPool::getOrAllocateDescriptorSet will handle this
retainResource call regardless of cache hit or miss.
Bug: b/372268711
Change-Id: I11e47ab9e56a24eb68a0231c5e553ab6b8833c82
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5921640
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
3e8d09a1
|
2024-10-08T17:01:02
|
|
Vulkan: Enable FRAGMENT_SHADER_FRAMEBUFFER_FETCH_MRT_ARM
ANGLE correctly handles MRT with gl_LastFragColorARM, this property
should have been set to true.
Bug: b/269233744
Change-Id: Iaa554df57292fcc5c12281e2e098b6451f63d5e6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5916770
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Cody Northrop <cnorthrop@google.com>
|
|
|
ba292370
|
2024-10-08T19:58:29
|
|
Vulkan: Disable imageless framebuffers on Qualcomm and PowerVR
* Qualcomm with imageless framebuffers, vkCreateFramebuffer
loops forever
* PowerVR with imageless framebuffer spends enormous amounts of time
in framebuffer destruction and creation. ANGLE doesn't cache
imageless framebuffers, instead adding them to garbage collection,
expecting them to be lightweight.
Tests: dead_cells, chrono_trigger, batman_telltale
Bug: b/369693310, b/372273294
Change-Id: I43714154661c4415ec10511ed096da7c23a9a8ef
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5918740
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Auto-Submit: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Cody Northrop <cnorthrop@google.com>
|
|
|
878e1c92
|
2024-10-07T16:34:54
|
|
Vulkan: Fix line-loop draw arrays after elements
Currently, when drawing line-loop arrays in Vulkan, an index
buffer is created to assist in drawing (since Vulkan does not
natively include a line-loop draw mode). However, when LL array
draw calls are mixed with non-LL element draw calls, it can lead
to some rendering issues due to the fact that the proper index
buffer is not obtained.
* In VertexArrayVk::handleLineLoop(), if the cached indices are the
same as the last draw, the same index buffer is obtained from the
LineLoopHelper object.
* (Using the newly added getCurrentIndexBuffer())
* Added unit test in which a triangle element draw is places between
two LL array draws. Before the fix, the second LL draw would result
in an incorrect line draw.
Bug: chromium:40911000
Change-Id: Ibba9a0cb2b77a2b6ae2c1e9230afe3d16b70cb63
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5908694
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
166b72c9
|
2024-09-30T19:07:26
|
|
GL_ANGLE_blob_cache implementation.
Bug: chromium:370538323
Change-Id: Ic51a951e78b48b315e36f518bcc39ff2d54660a6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5900761
Reviewed-by: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
b3d85cce
|
2024-09-30T14:28:35
|
|
Vulkan: Consolidate write colorspace override states
ColorspaceState struct is now used to cache write colorspace related
states to determine the colorspace of Vulkan draw image views.
ImageViewHelper methods are called during initialization and when
colorspace related states are toggled dynamically which in turn process
these states and determine the final write colorspace.
We can now fully support rendering to EGLImages, with colorspace
overrides, via texture or renderbuffer EGLImage targets
Bug: angleproject:40644776
Tests: ImageTest*Colorspace*Vulkan
MultithreadingTestES3.SharedSrgbTextureMultipleContexts*Vulkan
ReadPixelsPBOTest.SrgbUnorm*Vulkan
Change-Id: I2be2cd3b5b2b4ac8ecb803c34cde2b846cbd1cbe
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5901256
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
|
|
|
b38cc7fa
|
2024-09-30T12:43:09
|
|
Vulkan: Consolidate read colorspace override states
ColorspaceState struct is now used to cache read colorspace related
states to determine the colorspace of Vulkan read image views.
ImageViewHelper methods are called during initialization and when
colorspace related states are toggled dynamically which in turn process
these states and determine the final read colorspace.
Bug: angleproject:40644776
Tests: ImageTest*Colorspace*Vulkan
SRGBTextureTest.SRGB*TextureParameter*Vulkan
SRGBTextureTestES3.SRGBDecodeTexelFetch*Vulkan
Change-Id: I16b3666cd80865936b826dc0738fc9210dabeda9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5901255
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
|
|
|
605c2f85
|
2024-09-27T12:00:58
|
|
Vulkan: Bugfix in FramebufferVk::blit(...)
If any color attachment of GL_DRAW_FRAMEBUFFER is already in use in the
currently started renderpass, don't reuse the renderpass for blit.
Bug: angleproject:40644776
Tests: Texture2DTestES3.UnitTest_DMSAA_dst_read*
Change-Id: Ib4266fc65d0ac7151f362f263e9e706adb1231ce
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5895158
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
f9709279
|
2024-09-20T16:02:56
|
|
CL/Vulkan: Add support for printf builtin processing
The support for printf builtin in clspv enabled through the SPIR-V
non-semantic clspv reflection instructions - PrintfInfo and
PrintfBufferStorageBuffer [1]. The printf buffer is setup with a
separate descriptor layout and the pipeline layout is updated
accordingly.
Also, printf is enabled as default option now for clspv.
[1]: https://github.com/KhronosGroup/SPIRV-Registry/blob/master/nonsemantic/NonSemantic.ClspvReflection.html
Bug: angleproject:369724757
Change-Id: I20b245eb0fea69941bd1aeb42534f8b729ec17e8
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5893958
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
b61f9f9e
|
2024-10-04T11:07:23
|
|
Vulkan: Add operator<< for descriptorSet for debugging
Right now it is using streamOut() function which is hard to use with
WARN(). This replaces the streamOut function with standard c++
operator<< so that we can use in WARN()/INFO() along with other logs for
debugging.
Bug: b/368566032
Change-Id: Iec98b4c59f360cbbfb8fbdd85d5d1150fcca8f4a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5908773
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
d147a2ca
|
2024-10-04T10:45:55
|
|
Vulkan: release descriptorSets from TextureVk::refreshImageViews
When texture has to refresh its imageViews, the existing descriptorSet
cache will become obsolete (because the cache key contains view's
uniqueSerial). This CL proactively releases all cache so that we don't
keep these invalid cache any more.
Bug: b/368566032
Change-Id: I82a221b8b8fd072ee3a52b11264addba385526d5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5908772
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
eb4eaea9
|
2024-10-03T17:15:21
|
|
Vulkan: Improve SharedCacheKeyManager::addKey performance
This function walks a vector of keys. When there are many keys this
could be slow. Also when we have to grow the vector size, it involves
memory reallocation which means copy the data from old storage to new
storage. This CL changes mSharedCacheKeys to use std::deque instead of
vector which solves storage reallocation problem. It also adds
angle::BitSet64<64> to track all available (i.e., empty) slots in
mSharedCacheKeys so that we don't have to loop most of time. You only
loop all keys once to find all empty slots and then subsequent addKey()
call will be O(1) until all empty slots are used.
Bug: b/368566032
Change-Id: I4d32b461761f1cd64380f5527883b84357bb44c1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5908690
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
b16d105f
|
2024-10-03T10:25:32
|
|
Remove Desktop GL front-end support
For Desktop GL applications, please use Zink!
Bug: angleproject:370937467
Change-Id: Ie734634bb62a2e98c80e1b32d8b3d34624da3c04
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5905428
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
6024e9c0
|
2024-10-02T08:01:43
|
|
Manual roll VK-GL-CTS from 65470ff2e321 to 179dd9f858f0 (27 revisions)
https://chromium.googlesource.com/external/github.com/KhronosGroup/VK-GL-CTS.git/+log/65470ff2e321..179dd9f858f0
2024-09-30 lorenzo@khronosgroup.org Fix build
2024-09-30 mark@igalia.com Support standalone executable builds on
Android
2024-09-30 javed@igalia.com Add tests for draw count clamping with null
index buffer
2024-09-30 piotr.byszewski@mobica.com Capture replay tests for
VK_EXT_descriptor_buffer
2024-09-30 javed@igalia.com Add tests vkCmdDrawIndexedIndirectCount draw
count clamping
2024-09-30 rgarcia@igalia.com Add R8 and R16 large image tests in host
image copy test group
2024-09-30 ziga@lunarg.com Add missing linked shader object test case
2024-09-30 tapani.palli@intel.com Fix issues with GLX reset notification
strategy
2024-09-30 kamil.goras@mobica.com Split
KHR-GLES32.core.tessellation_shader.vertex.vertex_spacing
2024-09-30 jimblackler@google.com Handle resize events in EGL bounding
box tests.
2024-09-30 lorenzo@khronosgroup.org Merge vk-gl-cts/vulkan-cts-1.3.9
into vk-gl-cts/main
2024-09-30 lorenzo@khronosgroup.org Merge remote-tracking branch
'vk-gl-cts/dev/VK_EXT_depth_clamp_control' into main
2024-09-30 lorenzo@khronosgroup.org Update Vulkan spec to 1.3.296
2024-09-30 lorenzo@khronosgroup.org Merge remote-tracking branch
'vk-gl-cts/dev/VK_KHR_compute_shader_derivatives' into main
2024-09-20 lorenzo@khronosgroup.org Update SPIRV-Tools, SPIRV-Headers,
glslang, Vulkan-Docs, Vulkan-ValidationLayers external sources
2024-09-19 piotr.byszewski@mobica.com Check limits before running
mapping_to_large_index DRLR test
2024-09-19 ziga@lunarg.com Test pipeline without render pass or
VkPipelineRenderingCreateInfo
2024-09-19 piotr.byszewski@mobica.com Remove invalid dynamic rendering
local read tests
2024-09-19 piotr.byszewski@mobica.com Fix pipeline binary graphics tests
2024-09-19 piotr.byszewski@mobica.com Fix pNext chain loop in FSR tests
2024-09-19 kamil.goras@mobica.com Split KHR-Single-GL46.arrays_of_arrays
2024-09-19 marcin.hajder@mobica.com Port KC-CTS tests to VK-GL-CTS
(glGetUniform), PART 11
2024-09-19 marcin.hajder@mobica.com Port KC-CTS tests to VK-GL-CTS
(scissor_blit), PART 12
2024-09-19 marcin.zajac@mobica.com Suballocated DRM images
2024-09-19 piotr.byszewski@mobica.com Test feedback loop with dynamic
rendering local read
2024-09-19 lorenzo@khronosgroup.org Merge vk-gl-cts/vulkan-cts-1.3.9
into vk-gl-cts/main
2024-09-18 jimblackler@google.com Fix error in test logging in all
CtsDeqpTestCases
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/vk-gl-cts-angle-autoroll
Please CC angle-team@google.com,syoussefi@google.com on the revert to
ensure that a human
is aware of the problem.
To file a bug in ANGLE:
https://bugs.chromium.org/p/angleproject/issues/entry
To report a problem with the AutoRoller itself, please file a bug:
https://issues.skia.org/issues/new?component=1389291&template=1850622
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md
Bug: None
Change-Id: Idbc50f78119327844e1488db25d58c1c588f9ff6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5903935
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
b5d548bb
|
2024-09-20T17:06:54
|
|
CL/Vulkan: Update map interface for CLMemoryVk
Clean up the map interface of the CLMemoryVk to account for any offsets.
Bug: angleproject:369724757
Change-Id: I12c7bf2f7f94cd7fbf57a8ee10748d2275a2ba70
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5893956
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
0a452697
|
2024-09-19T16:33:26
|
|
CL/Vulkan: Enable support for multiple descriptor set handling
The clspv transcompiler can generate multiple descriptor sets depending
on the kernel e.g. printf storage buffer is setup in a separate
descriptor set, and kernel arguments in a separate descriptor set. This
change enables setting up of multiple descriptor sets and appropriately
bind them.
Also, creation of descriptor set and pipeline layout are deferred to
kernel enqueue time as it is quite possible that kernels are built in a
batch before their use.
Bug: angleproject:369724757
Change-Id: I80eb93f4a3b8afc7461c299cc283526cc4b872fa
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5893955
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
371539c3
|
2024-09-19T16:26:44
|
|
CL/Vulkan: Move descriptor set and pipeline layout cache to context
Move the descriptor set and pipeline layout caches to context so that
cache is shared across programs.
Bug: angleproject:369724757
Change-Id: I517d0cc712adb3f8b91d68e5bca8d3ac3ddb39ff
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5893954
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
435bd0a9
|
2024-10-02T16:37:06
|
|
ssci: use canonical date format
Bug: chromium:370881022
Change-Id: Ie8a310b11517fd1be7f95f246bcb9cfe48d95791
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5903336
Auto-Submit: Jiewei Qian <qjw@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
cc44090d
|
2024-09-18T12:28:53
|
|
Vulkan: Add an extra descriptor set index
In the case of CL, the clspv transcompiler can generate upto four
descriptor set indices, so add an extra index to vk::DescriptorSetIndex.
Also, adding aliases for CL specific naming.
Bug: angleproject:369724757
Change-Id: I45ef8a6d9246c7863ebc6edf08479bc7c661c151
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5893953
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
|
|
0040cda1
|
2024-09-25T18:16:36
|
|
Vulkan: Invalidate host visible non-coherent buffers on mapping
We can not trust the cache during CPU readback when the buffer
memory type is non-coherent.
Bug: b/366134076
Change-Id: I89920cfa468ee0be0feb607fea9d60bc0732191f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5890707
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Auto-Submit: Imran Ziad <imranziad@chromium.org>
|
|
|
67a5ea45
|
2024-09-23T16:09:12
|
|
Vulkan: Fix the error from multiple lineloop draws
Since Vulkan does not support line-loop draws natively, such a
draw call requires the conversion of the related buffers to prepare
them for this operation. For glDrawElementsIndirect(), the index
and the indirect buffers would need conversion.
However, what currently happens in this case is that the original
buffer pointer is overwritten after the conversion, removing the
link to the original buffer. Therefore, if there is a second line-loop
call just after the first, it will try to use the converted buffer as
the new source, which leads to errors due the buffer already being in
use.
The index buffer for the draw is bound when the related dirty bit is
handled. Therefore, instead of using the draw index buffer directly
for handling the line-loop scenario, we can use the index buffer in
the form of a local pointer passed between functions. Then, in order
to reconcile line-loop with the other cases, the draw index buffer is
set just before setting up the indexed draw.
* Functions handling line-loop draws do not modify the element array
buffer in VertexArrayVk directly, but use local buffer pointers to
pass the current element array pointer to further processing and
drawing.
* Added mCurrentElementArrayBuffer for ContextVk to be bound to the
index buffer to used for draw instead of the one from its vertex
array object.
* Before the indexed draw, mCurrentElementArrayBuffer is set to the
last destination index buffer.
* Added unit test that makes a line-loop draw and then a non-LL call
using the same element array.
Bug: angleproject:360758685
Change-Id: I6d6328f6326c1a1f9f80e5ef346aa077c867d344
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5878764
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
Reviewed-by: Charlie Lao <cclao@google.com>
|
|
e06b07a9
|
2024-09-27T18:17:16
|
|
Vulkan: populate ycbcr conversionDesc for yuv VkFormats
For this case, the external format has a vulkan equivalent VkFormat and
is YUV, so conversionDesc needs to be populated.
YcbcrConversionDesc::update() expects externalFormat to be 0 in this
case since it has an equivalent.
Bug: b/368429977
Bug: b/352595520
Change-Id: Ifdae7a539555d78bbbf9d11cdbb946cb815c7709
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5896250
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
7c811715
|
2024-09-25T11:09:44
|
|
Vulkan: fix crash when clearing stencil with ClearBuffer
Follow up to [1] which fixed a crash with glClear, but the bug remained
with glClearBufferiv. This change refactors the "is stencil write
masked out" query to always take the framebuffer's stencil bit count
into account (practically always 8), which also happens to make the rest
of the code checking this query more accurate in the presence of
nonsense masks where the bottom 8 bits are 0.
[1]: https://chromium-review.googlesource.com/c/angle/angle/+/3315158
Bug: chromium:40207259
Bug: angleproject:42266334
Change-Id: I68a6b0b75c67ed2cdc8c4d03b243efe5495efce1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5889788
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>
|
|
|
994bbbfc
|
2024-09-23T23:30:10
|
|
Vulkan: Don't require renderability in AHBs
If the AHB usage includes GPU_FRAMEBUFFER, in practice it should be
impossible for the corresponding Vulkan format to not be renderable.
However, to guard against potential driver bugs, renderability is still
checked if the AHB usage includes that bit.
Bug: b/368429977
Change-Id: Iae23432c369cd933f14c143f67c08720d76009fc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5881904
Reviewed-by: mohan maiya <m.maiya@samsung.com>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
2af09236
|
2024-09-24T13:55:04
|
|
Vulkan: Enable monolithic pipelines on Intel Windows
The hang issue is fixed and not reproduced on Intel
Windows drivers 101.5379 and later.
Bug: angleproject:369043994
Change-Id: Ic395ca47e65fa6467baf09e54b7e24c0d7d71ad3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5885989
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
eaffa034
|
2024-09-24T20:56:04
|
|
Revert "Vulkan: Consolidate colorspace override states"
This reverts commit bffcd235ba6c031603d798daaa98f1cf9a3f3e46.
Reason for revert: Breaks Android test `org.skia.skqp.SkQPRunner#UnitTest_DMSAA_dst_read`. Details:
https://b.corp.google.com/issues/369388539.
Original change's description:
> Vulkan: Consolidate colorspace override states
>
> ColorspaceState struct is now used to cache colorspace related states
> and used to determine the colorspace of Vulkan image views.
> ImageViewHelper methods are called during initialization and when
> colorspace related states are toggled dynamically which in turn process
> these states and determine the final read and write colorspaces.
>
> We can now fully support rendering to EGLImages, with colorspace
> overrides, via texture or renderbuffer EGLImage targets
>
> Bug: angleproject:40644776
> Tests: ImageTest*Colorspace*Vulkan
> MultithreadingTestES3.SharedSrgbTextureMultipleContexts*Vulkan
> SRGBTextureTest.SRGB*TextureParameter*Vulkan
> SRGBTextureTestES3.SRGBDecodeTexelFetch*Vulkan
> ReadPixelsPBOTest.SrgbUnorm*Vulkan
> Change-Id: I1cc2b5bd834b519b83deab4d80a2fcaabeb271d6
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5841290
> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
> Reviewed-by: Charlie Lao <cclao@google.com>
> Commit-Queue: mohan maiya <m.maiya@samsung.com>
Bug: angleproject:40644776
Change-Id: I5bf6cf2ed0c8ec22fc02d8c3da92673ee85fe002
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5888506
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
f5f419ec
|
2024-09-04T20:22:20
|
|
Vulkan: Add verify-restore in CompressAndStorePipelineCacheVk()
Change fixes following problem:
When blob cache is full, but still allows to store the current
pipeline cache data, storing next chunk may trigger eviction of
already stored items. Depending on the blob cache implementation,
eviction process may choose to evict chunks from the current pipeline
cache data. As the result: blob cache will not contain all chunks.
Solution:
The `StorePipelineCacheVkChunks()` function now check what chunks
exist in the blob cache (if `lastNumStoredChunks` parameter is positive)
and restores any missing chunks if their number is less than
`lastNumStoredChunks`.
First call to `StorePipelineCacheVkChunks()` passes zero (0) to the
`lastNumStoredChunks`, which will skip the check and store all chunks
unconditionally.
If "verifyPipelineCacheInBlobCache" feature is enabled, the
`StorePipelineCacheVkChunks()` will be called again to restore possibly
missing chunks.
If "hasBlobCacheThatEvictsOldItemsFirst" feature is disabled,
the function will be called repeatedly in a loop (opposed to only once),
until all chunks are present or until number of missing chunk stops
decreasing.
Bug: angleproject:4722
Change-Id: Ibbbdb2f878d6aafd7ab5586fc654e2b9be900a9a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5837042
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
|
|
|
a6ec0bb9
|
2024-09-23T10:50:28
|
|
Vulkan: Fix recursion in ensurePipelineCacheInitialized()
If the initial data was loaded from the blob cache, this function called
getPipelineCacheSize() to initialize mPipelineCacheSizeAtLastSync.
However, that function itself lead back to
ensurePipelineCacheInitialized(), causing a recursion. This was
previously undiscovered as the conditions that would lead to the global
pipeline cache needing a lock were yet to materialize in the wild.
Since the preferMonolithicPipelinesOverLibraries feature was made more
widely enabled in [1] and consequently a race condition was fixed in
[2], this recursion was discovered as a deadlock due to the mutex use.
While this change avoids the recursion, it simultaneously optimizes the
syncPipelineCacheVk() function by making sure the lock is taken once
intstead of twice when retrieving the pipeline cache size and
subsequently data. This also avoids a previously encountered race
condition where the pipeline cache was modified in between the two
queries and VK_INCOMPLETE was returned from the second call.
[1]: https://chromium-review.googlesource.com/c/angle/angle/+/5870466
[2]: https://chromium-review.googlesource.com/c/angle/angle/+/5872715
Bug: angleproject:42265839
Change-Id: Ic682b3d20ec4411ba180b3bafb807fdde8166d5b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5883153
Commit-Queue: Roman Lavrov <romanl@google.com>
Reviewed-by: Steven Noonan <steven@uplinklabs.net>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Roman Lavrov <romanl@google.com>
|
|
fdec6935
|
2024-09-20T16:29:03
|
|
Workaround supportsSurfaceMaintenance1 on Linux with llvmpipe
Presence of multiple ICD appears to confuse the loader and we end up
with support reported but vkGetPhysicalDeviceSurfaceCapabilities2KHR
not actually working (see bug for more detials)
Bug: angleproject:368647924
Change-Id: Idc813b99f873643d067cc4b4e38f1ed5249500b4
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5875015
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Roman Lavrov <romanl@google.com>
|
|
|
1e74ce33
|
2024-09-17T14:44:36
|
|
Reland "Vulkan: Prefer monolithic pipelines everywhere"
This reverts commit 1503a10737e28ac9d1bb98c2f3e369880c63ca43.
Reason for revert: Fixed threading + undo for buggy drivers
Original change's description:
> Revert "Vulkan: Prefer monolithic pipelines everywhere"
>
> This reverts commit e9ab557c744c29088174b6c7bc93692fe7f32dbd.
>
> Reason for revert: kills Windows Intel UHD 770 machines
> http://anglebug.com/42265839#comment34
>
> Original change's description:
> > Vulkan: Prefer monolithic pipelines everywhere
> >
> > Apparently, every vendor prefers this, and
> > graphicsPipelineLibraryFastLinking has no bearing on the driver's
> > preference.
> >
> > Bug: angleproject:42265839
> > Change-Id: Ied8d82e0a9e08fe5efe3f625e2612764f67f9768
> > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5861024
> > Commit-Queue: Yuxin Hu <yuxinhu@google.com>
> > Reviewed-by: Yuxin Hu <yuxinhu@google.com>
> > Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
>
> Bug: angleproject:42265839
> Change-Id: I95fa61b2242684b63adc2841b777bde923e34d41
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5868055
> Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Bug: angleproject:42265839
Change-Id: I007a3efc69e0edc040b0feddcd84e191f68ded9d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5870466
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
bffcd235
|
2024-09-13T14:58:00
|
|
Vulkan: Consolidate colorspace override states
ColorspaceState struct is now used to cache colorspace related states
and used to determine the colorspace of Vulkan image views.
ImageViewHelper methods are called during initialization and when
colorspace related states are toggled dynamically which in turn process
these states and determine the final read and write colorspaces.
We can now fully support rendering to EGLImages, with colorspace
overrides, via texture or renderbuffer EGLImage targets
Bug: angleproject:40644776
Tests: ImageTest*Colorspace*Vulkan
MultithreadingTestES3.SharedSrgbTextureMultipleContexts*Vulkan
SRGBTextureTest.SRGB*TextureParameter*Vulkan
SRGBTextureTestES3.SRGBDecodeTexelFetch*Vulkan
ReadPixelsPBOTest.SrgbUnorm*Vulkan
Change-Id: I1cc2b5bd834b519b83deab4d80a2fcaabeb271d6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5841290
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
|
|
|
167b9e8d
|
2024-09-18T21:51:38
|
|
Vulkan: Fix pipeline cache store vs monolithic pipeline race
The thread that creates monolithic pipelines needs to hold the pipeline
cache lock, as well as the thread that stores the pipeline cache
contents to the blob cache.
Bug: angleproject:42265839
Change-Id: I17cf9d2bb3f27d531f368003cb4ee00007a464fa
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5872715
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
1bed7fdd
|
2024-09-13T15:29:07
|
|
CL/VK: Fix missed PushConstantRegionGroupOffset
Adding missing reflection instruction to parser:
NonSemanticClspvReflectionPushConstantRegionGroupOffset
And add related PC routine during NDRangeKernel.
Bug: angleproject:366412385
Change-Id: I2ffa48dee46a23840390312f20e7103f68daa0c5
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860310
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
cce1497c
|
2024-06-07T15:03:07
|
|
CL/Vulkan: regionOffset as GWO for uniform
clspv states:
region_offset: the sum of the global ID
offset into the NDRange for this uniform
region and the global offset of the NDRange
Since we currently do not support non-uniform at
the moment, we just treat the
PushConstantRegionOffset as the global work offset.
Bug: angleproject:366412385
Change-Id: Ifd9953f21cd7826198b2cfe7d30ddec68ea2c7a0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860309
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
741e5355
|
2024-06-07T16:39:53
|
|
CL/Vulkan: Add missing PushConstantNumWorkgroups
Adding support for:
NonSemanticClspvReflectionPushConstantNumWorkgroups
Bug: angleproject:366412385
Change-Id: Id4cfc1ee456d3a7f6d058f3f4a6a580381b2d471
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860308
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
8eecc8e9
|
2024-05-30T13:30:08
|
|
CL/Vulkan: Update CL_DEVICE parameters
Update device fields for the following:
- Change address bits from 64 -> 32
- Report OpenCL C version to 1.2
- Change non-uniform wgs support to false
We either did not support nor take full advantage
of these features, thus leading to the omission
or change/downgrade.
Bug: angleproject:366412386
Change-Id: Ia6b361d0285d5a1a88013a49fd747a3d03db9655
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860307
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
cb9d52f8
|
2024-09-11T15:26:49
|
|
Vulkan: Add Pipeline blob cache per chunk CRC check
Each chunk now have it's own CRC value, that is calculated starting
from beginning of the data up to the end of the chunk. This way last
chunk will have CRC of the entire data, avoiding calculating it
separately.
The `GenerateCRC32()` was splint into `InitCRC32()` and `UpdateCRC32()`
function, which supports calculating running CRC32.
Calculating CRC this way has some performance hit, when using many
small chunks. Increasing chunk size to at least 1 MB almost negates any
overhead from running `UpdateCRC32()` multiple times.
On S921B with locked frequencies, calculating CRC for 89 MB of data
using single call took 7.13 ms and 13.34 ms, when using multiple calls.
However even this difference is negligible, because compressing 291 MB
of data to 89 MB took 9.22 seconds.
Bug: angleproject:4722
Change-Id: Ic6f3245b3aa7319ac13502d2c43e1a084a9bc8b2
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5854709
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
8469debb
|
2024-09-04T20:04:10
|
|
Vulkan: Add dual slots in CompressAndStorePipelineCacheVk()
Change fixes following problem:
Currently, each call to `CompressAndStorePipelineCacheVk()` stores
chunks in order, starting from 0. This overrides previously stored
chunks. In case of app termination (kill) in the middle of this
process, the entire cache data will be corrupted, since it will
partially contain chunks from the new and old caches.
Solution:
In order to fix this problem, this change introduces `slotIndex` into
the `chunkCacheHash` calculation. Slot index is managed by
`vk::Renderer::getNextPipelineCacheBlobCacheSlotIndex()` method, which
will alternate between 0 and 1 when "useDualPipelineBlobCacheSlots"
feature is enabled, and always 0 otherwise.
Additionally, chunk storing order is reversed: last chunk is stored
first and the first (0 chunk) - last. This is done because 0 chunk is
the first that is loaded in `GetAndDecompressPipelineCacheVk()` and used
as indication that there is data in the cache. Writing it last, ensures
that other chunks will be also available.
When "useDualPipelineBlobCacheSlots" is enabled, each call to
`CompressAndStorePipelineCacheVk()` will use slot index opposed to the
slot that is stored in the cache, avoiding damaging existing data. After
writing all chunks for a brief moment there may be 2 instances of the
data. However, data for the previous slot will be immediately erased
(by writing 1/0-sized blobs) starting from the 0 chunk.
To control if erasing of old pipeline cache data will be erased by using
0-sized or 1-sized blobs blobs, added
`useEmptyBlobsToEraseOldPipelineCacheFromBlobCache` feature.
The `GetAndDecompressPipelineCacheVk()` function will iterate over each
available slot index checking only 0 chunk until data is found.
In case of the OpenCL API, features will always have following values:
- "useDualPipelineBlobCacheSlots" -> false
- "useEmptyBlobsToEraseOldPipelineCacheFromBlobCache" -> true
Note: this solution requires 2X pipeline cache size space in the blob
cache to work as expected, otherwise it will exacerbate other problem:
When blob cache is full, but still allows to store the current
pipeline cache data, storing next chunk may trigger eviction of
already stored items. Depending on the blob cache implementation,
eviction process may choose to evict chunks from the current pipeline
cache data. As the result: blob cache will not contain all chunks.
The above problem will be addressed in the follow up CL.
Bug: angleproject:4722
Change-Id: I2920bc3d89263280cdfe0466446fca26415e2b25
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5756576
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Igor Nazarov <i.nazarov@samsung.com>
|
|
|
44b6203c
|
2024-09-02T16:23:44
|
|
Add new functions defined in VK_KHR_Synchronization2 extension
Add vkCmdPipelineBarrier2 and vkCmdWriteTimeStamp2.
The other functions introduced in VK_KHR_Synchronization2
are not being used at the moment.
Bug: b/356985874
Change-Id: I4f8b432aee9b11debcb8abe8b3089f88ad7a069b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5832709
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
|
59f1e537
|
2024-09-03T17:56:47
|
|
CL/VK: Add ArgumentWorkgroup handling
Each OpenCL kernel can have
ArgumentWorkgroup args - which are
treated as spec constants.
Bug: angleproject:366415134
Change-Id: I2761010610e5b991e959006e7b78f1241a687960
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860304
Commit-Queue: Austin Annestrand <a.annestrand@samsung.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
7bdd5f77
|
2024-09-18T12:51:25
|
|
Vulkan: Remove syncval suppressions for fixed bugs
Bug: angleproject:42265220
Change-Id: I282f77330867682eee121f67d5357658bf375844
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5873529
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
|
e38d25b1
|
2024-06-21T18:22:32
|
|
Vulkan: Implement EXT_clear_texture
* Added new functions to TextureVk to clear the image.
* clearImage()
* clearSubImage()
* Both implemented via clearSubImageImpl(), with the former a
special case of the latter.
* For multisample or renderable images, stagePartialClear() from
ImageHelper is called to add the update.
* For single-sampled non-renderable images, a buffer is filled with
the pixel data and applied to the image as a buffer update.
* Added new update type: ClearPartial
* Used for renderable textures. This includes multisample textures.
* LOAD_OP_CLEAR is used in a render pass to perform the clear.
* UtilsVk::clearTexture()
* (Uses ClearTextureParameters)
* Uses the following functions to get the VkClearValue from the
input data and format:
* GetVkClearColorValueFromBytes()
* GetVkClearDepthStencilValueFromBytes()
* ClearPartial updates can also be superseded and removed similar to
Buffer updates.
* Updated UtilsVk::startRenderPass() to accept a VkClearValue* as an
input arg. If used, the render pass will use LOAD_OP_CLEAR.
* Enabled the feature "clearTextureEXT" on Vulkan.
* Added new unit tests in ClearTextureEXTTest for various formats and
pixel sizes.
* Added related multisample tests in FramebufferTest.cpp.
* FramebufferTest_ES31.ClearTextureEXT*
* Disabled some of the new tests failing using OpenGL.
* Disabled stencil-only-related tests on Pineapple.
Bug: angleproject:42266869
Change-Id: I89c631d68a4ed63d9991abe1783333255ade20dd
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5778348
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
7c6f1c05
|
2024-08-20T17:50:00
|
|
Vulkan: Enable UtilsVk::startRenderPass() for DS
Currently, UtilsVk::startRenderPass() assumes usage for color ops
and only uses color aspect flags and layouts. This CL will add the
option to use it for depth/stencil as well.
* Added the aspect flags to the args of startRenderPass()
* For the existing usage of this function, which use the color bit,
the image layout "ColorWrite" is used as before.
* For depth/stencil usage, the layout to be used therein will be
DepthWriteStencilWrite.
* The follow-up CL will use this.
Bug: angleproject:42266869
Change-Id: I135df3b91064b6e711e6109e9e7b734bc6361396
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5801657
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
7e462c22
|
2024-09-17T15:32:41
|
|
WGPU: Implement SetUniform() enough so GetUniform() works
Lays out a shadow buffer for basic uniforms per-shadertype
in std140, which is close to matching WGPU's layout. This
does not actually pass the buffer to WGPU as a uniform
buffer. GetUniform() just reads from the shadow buffer.
This is copied from the VK backend and so some code is
deduplicated.
Bug: angleproject:42267100
Change-Id: I727dc9e09a7ccabbb617f148dd68590469883b07
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5867444
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Reviewed-by: Liza Burakova <liza@chromium.org>
|
|
|
dfa516ce
|
2024-09-12T15:25:08
|
|
Vulkan: Fix chunkIndex is not 16 bits when computing key
Replace `uint8_t` with `size_t` in `ComputePipelineCacheVkChunkKey()`.
This fixes the bug and also makes type consistent with the rest of the
code.
Also `CacheDataHeader::setData()` made consistent with `getData()`.
Bug: b/246683126
Change-Id: Ib1f8b680622c3d4f912d8a21ed8b740c3d17483e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5854707
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
b98e6bc0
|
2024-09-12T00:00:00
|
|
Translator: Add EXT_texture_shadow_lod functions
Added translator support for functions defined
in the GL_EXT_texture_shadow_lod extension.
Trivially enabled on GL and Vulkan backends.
Bug: angleproject:365066518
Bug: angleproject:365108862
Change-Id: Ie1fc8f50e321a559d4506c479d9cebbb48802091
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5867416
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
6e0ae899
|
2024-09-17T12:12:03
|
|
Vulkan: Suppress more rasterizer-discard VUs due to VVL bugs
Bug: angleproject:42266678
Change-Id: I7321c28362d1a2be529b9d173c8429c757c26fd8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5869444
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
dbdc9551
|
2024-09-16T10:19:47
|
|
Vulkan: Let asyncCommandBufferReset control garbage cleanup
So that people can toggle the flag to compare perf/power difference with
async thread doing garbage clean up AND command buffer reset. This also
renames feature flag asyncCommandBufferReset to
asyncCommandBufferResetAndGarbageCleanup to reflect the implementation.
Bug: b/255411748
Change-Id: Id459e6f4dc81ec76b6c0c2dba0db46041ea6ae8a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5867389
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>
|
|
|
72d01332
|
2024-05-23T16:34:14
|
|
CL/Vulkan: Add missing EnqueuedLocalSize PushConst
Additionally, we move down "processKernelResources"
after "getOrCreateComputePipeline" since new push
constant needs LWS value and we might modify
ndrange's LWS after "getOrCreateComputePipeline".
Bug: angleproject:366412385
Change-Id: Ia287d667e0616256d1f9c039566b79aae1337154
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860306
Commit-Queue: Austin Annestrand <a.annestrand@samsung.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
9685cb96
|
2024-09-17T11:41:28
|
|
Vulkan: Clean up PackedClearValuesArray::store
The code was a bit confusing, raising question why there is check of
aspectFlags != VK_IMAGE_ASPECT_STENCIL_BIT. I believe this was
originally from
https://chromium-review.googlesource.com/c/angle/angle/+/2142711, and
then followed by
https://chromium-review.googlesource.com/c/angle/angle/+/2410366. The
code seems to having some special logic to handle packed depth stencil
clear value. This code may have been changed a bit. When I am looking at
current code base, I am not seeing good reason doing it this way. Caller
is handling the packed format by reading out depth value and pack them
together and then store. This CL replaces store/storeNoDepthStencil
pair to storeColor/storeDepthStencil.
Bug: b/167301719
Change-Id: I40cfca1e51654f5ddaf4b2e8460ae5a26c656f2b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5870921
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
89f9ea08
|
2024-06-18T10:23:22
|
|
CL/Vulkan: Fix missing init() for clLinkProgram
Fixing missed init() call/routine when creating
the CLProgramVk object during clLinkProgram.
Bug: angleproject:366411889
Change-Id: I13402b823746a59b3f0539830a6cde64b6201ae9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860305
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Austin Annestrand <a.annestrand@samsung.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
|
|
|
f2d9d755
|
2024-09-10T10:28:37
|
|
CL: Drop image_descriptor and format from image create
The image descriptor and format are part of the state and as such remove
them passing down through the create()
Bug: angleproject:365694588
Change-Id: I6e5c4595c0093f15df65f43ed607643cb7c2ae14
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5854004
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
ad4159f3
|
2024-06-20T12:26:32
|
|
Vulkan: Add utils for cl state to vk state
Add utils for translation of vk state from/to cl state in vk_cl_utils.
Bug: angleproject:365694588
Change-Id: I74f33416a2746048c0b74793b06bee612cf51eb3
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5854002
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
f8031fbe
|
2024-09-17T11:35:58
|
|
Vulkan: Unsuppress fixed VU errors
Bug: chromium:359904720
Bug: angleproject:362545033
Change-Id: I2721344843ffbddd6c56261359360473ad896a39
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5870469
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
|
f2eb8781
|
2024-09-14T16:16:53
|
|
Vulkan: Selectively enable emulateR32fImageAtomicExchange
Only emulate R32F imageAtomicExchange if shaderImageFloat32Atomics
feature is not supported
Bug: angleproject:42264071
Change-Id: I305ab88bf3ac918eff5d8c399f0ed02ec8c60c2d
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860814
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
1503a107
|
2024-09-17T10:46:29
|
|
Revert "Vulkan: Prefer monolithic pipelines everywhere"
This reverts commit e9ab557c744c29088174b6c7bc93692fe7f32dbd.
Reason for revert: kills Windows Intel UHD 770 machines
http://anglebug.com/42265839#comment34
Original change's description:
> Vulkan: Prefer monolithic pipelines everywhere
>
> Apparently, every vendor prefers this, and
> graphicsPipelineLibraryFastLinking has no bearing on the driver's
> preference.
>
> Bug: angleproject:42265839
> Change-Id: Ied8d82e0a9e08fe5efe3f625e2612764f67f9768
> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5861024
> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Bug: angleproject:42265839
Change-Id: I95fa61b2242684b63adc2841b777bde923e34d41
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5868055
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
|
1b4d6185
|
2024-09-12T09:18:46
|
|
Vulkan: Cleanup sRGB related code
Image and image view code is littered with sRGB related enums, even
in places that don't deal with sRGB. Remove sRGB related parameters
from initLayerImageView and getLevelLayerDrawImageView methods, which
now assume default values. Add dedicated methods that allow overriding
sRGB state values.
Also introduce ColorspaceState struct that consolidates all sRGB
related states, this will be used in follow up changes to track
and infer colorspace of image views
Bug: angleproject:40644776
Change-Id: Ifb366db48043e376f9ff6c30c852c44dd96562a1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860808
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Commit-Queue: mohan maiya <m.maiya@samsung.com>
|
|
|
e9ab557c
|
2024-09-14T22:11:18
|
|
Vulkan: Prefer monolithic pipelines everywhere
Apparently, every vendor prefers this, and
graphicsPipelineLibraryFastLinking has no bearing on the driver's
preference.
Bug: angleproject:42265839
Change-Id: Ied8d82e0a9e08fe5efe3f625e2612764f67f9768
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5861024
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
937c5dc8
|
2024-09-09T12:18:55
|
|
Vulkan: Make image{Read,Write} helper interface api agnostic
Removing ContextVk dependency on the imageRead/imageWrite helper utility
functions.
Bug: angleproject:42266971
Change-Id: I493e1fb11e8ae192f766c822cbee278c49c23bfe
Signed-off-by: Gowtham Tammana <g.tammana@samsung.com>
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5845197
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
6ee07959
|
2024-09-13T17:12:26
|
|
Remove the suppressed VVL error
https://github.com/KhronosGroup/Vulkan-ValidationLayers/issues/8523
is fixed. Remove the VVL error from the suppressed list.
Bug: b/365580001
Change-Id: Ib7a3819e740393306995df1eabec14e1b9878d7a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5860811
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
8c12874c
|
2024-09-12T11:26:29
|
|
vulkan: add double buffer swapchain for fifo mode.
Double buffer swapchain usually drop the performance.
But at B* the default fps cap for silver device still at 30fps.
This change is for reduce latency.
Bug: b/311022968
Change-Id: Ida4044f439bbe3f235d53f5d1d2f945533cbb094
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5858255
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
7dcfc32b
|
2024-09-12T16:25:28
|
|
Remove renamed VUID
VUID-VkVertexInputBindingDivisorDescriptionEXT-divisor-01870
was renamed to
VUID-VkVertexInputBindingDivisorDescriptionKHR-divisor-01870
in
https://github.com/KhronosGroup/Vulkan-ValidationLayers/commit/6adb436617e6c66fbc8b2a24996659f5c5bbb63f
Bug: angleproject:42266877, angleproject:42266639
Change-Id: I382625d33ef6728835d8e9db307d32b7910ac9b7
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5855187
Commit-Queue: Solti Ho <solti@google.com>
Reviewed-by: Solti Ho <solti@google.com>
Auto-Submit: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
|
|
|
c9d55051
|
2024-09-06T10:56:07
|
|
Vulkan: Consolidate dirtyRanges before vertex conversion
Detect two ranges overlap or are continuous and merge them. This reduces
number of dispatch calls as well as avoids redundant conversion.
Bug: b/357622380
Change-Id: I06b73a1e9fd573d79af985b247f4d66bf97f756e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5851642
Auto-Submit: Charlie Lao <cclao@google.com>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
958b657f
|
2024-09-10T20:34:57
|
|
Manual roll vulkan-deps from 725499142cb6 to 3b92cef97feb (32 revisions)
Manual roll requested by yuxinhu@google.com
Manual roll changes: suppress VUID-vkCmdDrawIndexed-Input-08734
until VVL issue angleproject:365580001 is addressed.
https://chromium.googlesource.com/vulkan-deps.git/+log/725499142cb6..3b92cef97feb
Also rolling transitive DEPS:
https://chromium.googlesource.com/external/github.com/KhronosGroup/glslang/+log/a496a34b439022750d41d2ba04fbbe416ef81c9a..12a17b7ce41436427e358608183100b1103274da
https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Headers/+log/efb6b4099ddb8fa60f62956dee592c4b94ec6a49..2a9b6f951c7d6b04b6c21fe1bf3f475b68b84801
https://chromium.googlesource.com/external/github.com/KhronosGroup/SPIRV-Tools/+log/e1782d6675b88225225e331a6318554d473c54db..d160e170d74ff45cb2a88dfb365bdfd896016f7c
https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-Loader/+log/c758bac8bf1580b5018adafd3a2ec709237b0134..1108bba6c97174d172d45470a7470a3d6a564647
https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-Utility-Libraries/+log/fbb4db92c6b2ac09003b2b8e5ceb978f4f2dda71..ea5774a13e3017b6d5d79af6fba9f0d72ca5c61a
https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-ValidationLayers/+log/af7b0a35d009b5ad6e0b280a5b81388608ebfe39..99de3c17fbc2db6b6da0347916c9e01a383c2758
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/vulkan-deps-angle-autoroll
Please CC angle-team@google.com,yuxinhu@google.com on the revert to
ensure that a human
is aware of the problem.
To file a bug in ANGLE:
https://bugs.chromium.org/p/angleproject/issues/entry
To report a problem with the AutoRoller itself, please file a bug:
https://issues.skia.org/issues/new?component=1389291&template=1850622
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md
Bug: angleproject:365580001
Tbr: yuxinhu@google.com
Change-Id: Iac9c8b6310d3ebb955dd09fd26b71d373e13b479
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5849280
Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
|
|
|
10d56e63
|
2024-09-06T10:56:07
|
|
Vulkan: use mEntireBufferDirty if range covers entire buffer
Detect the dirtyRange covers entire buffer and set mEntireBufferDirty
instead of add range into mDirtyRanges. This will get into a much
simpler code path that will not have any redundant conversion. This also
removes quite big portion of overlapped ranges cases without much of
overhead.
Bug: b/357622380
Change-Id: Iedaa3662a6fc52257e71d39ab75baddf6ad3e41b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5840476
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Charlie Lao <cclao@google.com>
|
|
|
6afcbd3e
|
2024-09-10T10:31:42
|
|
Vulkan: Disable dynamic rendering on ARM
Due to driver bugs.
Bug: b/356051947
Change-Id: Ic345496010b051ad054811a8c34bb3f816bbf258
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5851632
Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
|
941b3df3
|
2024-08-21T17:23:16
|
|
Vulkan: Add new threshold for CPU buffer subdata
Currently, with preferCPUForBufferSubData enabled, when the GPU is
busy, we use the CPU to copy subdata, which includes duplicating the
buffer. If the buffer is large, this can cause performance regression.
To avoid this, we should avoid using the CPU in cases that can lead to
such regressions.
* Enabled preferCPUForBufferSubData for all ARM devices.
* This feature is only used in the following cases:
* If the buffer is smaller than than 32K; or
* If the copy size is greater than 1/8 of the total buffer size.
* Significant performance improvement observed in several traces.
Bug: b/360118138
Change-Id: Ibc8e3de9b4e081c69457c85facba893283c8fb38
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5824347
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
|
|
|
53476d6f
|
2024-09-04T14:34:11
|
|
Vulkan: Do vertex conversion with fine grain dirtyRange
Right now when we do vertex conversion with multiple dirty ranges, we
are merging dirty ranges into single range and then
UtilsVk::convertVertexBuffer() is called for the merged dirty range. If
there is big gap between two ranges, the merged range could be very big.
This means we end up doing many unnecessary conversion. This CL tracks
individual dirty ranges and issues dispatchCompute for each dirty range,
thus minimize the unnecessary conversions.
On S24, this change further reduces TraceTest.gangstar_vegas frame time
from ~6.0ms to ~3.8ms, almost parity to native GLES's ~3.6ms.
Bug: b/357622380
Change-Id: Ia103f3963bdb5996ff3f95164c955a3e4f33f311
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5787633
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Yuxin Hu <yuxinhu@google.com>
|
|
|
fc65058c
|
2024-05-03T14:47:29
|
|
CL/Vulkan: Rework spec constant handling
Specialization constants can be combined into
a single hash map that's later iterated on
when we create the compute pipeline.
Bug: angleproject:364396920
Change-Id: I161356808ff0bd6a589f01854264210011bec512
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5834664
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Austin Annestrand <a.annestrand@samsung.com>
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
45ea2094
|
2024-09-06T10:36:07
|
|
Don't build api_dump for secondary abi on Android
32-bit builds of VK_LAYER_LUNARG_api_dump are extremely slow so skip
them when building for arm64.
Bug: b/42267058
Change-Id: I8c2d959d86a58537dafa518fb58267b38cda21f4
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/5842279
Commit-Queue: Roman Lavrov <romanl@google.com>
Reviewed-by: Roman Lavrov <romanl@google.com>
Reviewed-by: Mark Łobodziński <mark@lunarg.com>
|