ir::Value has a virtual Type() method. Implement this on the subclasses
that actually have a type.
Removed pointless memory usage on instructions that have no value.
Also un-getter some fields to match the other IR code.
Change-Id: Ibf545a582d5cbb820410c9aaec0312692e803373
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131745
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
d3d11 doesn't allow creating a constant buffer with other GPU
accelerated usage. This CL workarounds problem by creating two
buffers one for uniform buffer usage, one for other usage, and
copy content to uniform buffer when it is needed.
Bug: dawn:1755
Bug: dawn:1798
Bug: dawn:1721
Change-Id: I26bfee1cca2204f6464ba611872c490165e97f68
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/132020
Commit-Queue: Peng Huang <penghuang@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds support for the transient attachment feature and texture
usage flag to Vulkan, using the usage flag to allocate textures via
lazily-allocated memory if the latter is available.
Testing is covered by the
MultisampledRenderingWithTransientAttachmentTest now being run on
Vulkan due to transient attachments feature now being supported on
Vulkan.
Bug: dawn:1695
Change-Id: I45a04d21b1b6ea612086a368b9c77a0ff43e93f8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/130180
Reviewed-by: Loko Kung <lokokung@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Colin Blundell <blundell@chromium.org>
There's a few reasons for this change:
* Not all values have identifiers, and carrying redundant fields is
inefficent.
* Not all IDs will be integers - much like LLVM IR and SPIR-V, we will
likely want to disassemble with textual identifiers, so a uint32_t
is not ideal, and a std::string is even more bloat for each value.
* Transforms don't use identifiers, but instead raw pointers. We don't
want to encourage using IDs as they're simply a less-efficient way to
refer to values.
* This makes values consistent with types and flow-control blocks, as
they will both have their disassembly ID generated by the
disassembler.
The next step will be to add a hashmap to the module so that
pre-declared value names can be stored out-of-band.
Bug: tint:1718
Change-Id: I80beafc165f2bde54cc44a91015776926ca952b2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131740
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Emit the OpEntryPoint instruction with the pipeline stage. Interface
variables will be done later.
Emit OpExecutionMode instructions for the workgroup size and fragment
shader origin, depending on the pipeline stage.
This is a reland of commit 90789ea1f8
with compilation fixes for MSVC.
Bug: tint:1906
Change-Id: If5290df504ebc8925b3c52510acab123edfd1073
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131800
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Auto-Submit: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
When traverseing an accessor expression, stop as soon as we hit a
source object that has a constant value. This prevents us from trying
to emit expressions that have abstract types and no materialization
nodes.
Bug: chromium:1442551
Change-Id: I8296ae58e63624e647052cdf966dbff15630a4d8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/132040
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
This introduces the notion of FeatureLevel, currently consisting of
Core or Compatibility. Each AdapterBase is now constructed with the
FeatureLevel it supports.
When discovering PhysicalDevices, create an AdapterBase for each of
the FeaturLevels which that PhysicalDevice supports. For most of the
backends, this will mean Core and Compatibility, while OpenGL and
D3D11 support only Compatibility.
Bug: dawn:1796.
Change-Id: I828247ef43e2220805ccf6c08827aa5e2382a026
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/118240
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Stephen White <senorblanco@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
The problem is becasue a resource can not be set as input and
output at same time on device context. So we have to track used
slots and unset related slots before binding a group.
This CL also unset all affect slots when a render pass or compute
pass is end, so all related resources could be unref from the
device context.
Bug: dawn:1705
Change-Id: I597762ad8afa3b8df7139b0070f0b457d7319836
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131380
Commit-Queue: Peng Huang <penghuang@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
ToProgram() will most likely be conditionally compiled as a separate .h
and .cc file. With just one static method, ir::Converter doesn't make a
lot of sense to be a class. Just replace with a free-function.
Bug: tint:1902
Change-Id: I305bf9cb73082a94848ec5b94129feabdb9b6fe9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131620
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL fix an issue related to tiering the maxBufferSize and
maxStorageBufferBindingSize limits. In detail, this CL:
1. Fix the maxStorageBufferBindingSize tiers to [128MB, 1GB, 2GB-4,
4GB-4] instead of the original last two 2GB-1 and 4GB-1, holding the
guarantee that maxStorageBufferBindingSize is a multiple of 4 bytes.
2. Add a maxBufferSize tier 4GB, ensuring the guarantee that
maxStorageBufferBindingSize must be no larger than maxBufferSize in
case of it tiered to 4GB-4. Previously the largest maxBufferSize tier
is 2GB, making this guarantee broken after tiering.
3. Move the adapter limits normalization logics in adapter initializing
to Limits.cpp to allow unittest. Related unittests implemented.
4. Normalize tiered limits to ensure that tiered
maxStorageBufferBindingSize and maxUniformBufferBindingSize are no
larger than tiered maxBufferSize. Related unittests implemented.
Issue: dawn:1780
Change-Id: I4821f196fa89c7f18ebbf8e5e45df1c3268db895
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/130120
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Emit the OpEntryPoint instruction with the pipeline stage. Interface
variables will be done later.
Emit OpExecutionMode instructions for the workgroup size and fragment
shader origin, depending on the pipeline stage.
Bug: tint:1906
Change-Id: Ieeeda5f17da48a8cf0d3344d3b254542c7198cb9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131381
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: James Price <jrprice@google.com>
