| Age | Commit message (Collapse) | Author |
|---|
|
* Open existential on arguments after overload resolution.
* Fix.
* Update source/slang/slang-check-overload.cpp
Co-authored-by: ArielG-NV <159081215+ArielG-NV@users.noreply.github.com>
---------
Co-authored-by: ArielG-NV <159081215+ArielG-NV@users.noreply.github.com>
|
|
(#4977)
* Add a test to ensure extension does not override existing conformance.
* Fix doc.
* Update documentation.
* Fix doc.
* Add diagnostic test.
|
|
|
|
* Support specialization constants.
* Fix.
* Fix.
* Fix.
* Fix.
* Make sure specialization constants have names.
* Clean up and support the dxc [vk::constant_id] syntax.
* Fix.
* Fix.
* Fix.
|
|
* add slang-rhi submodule
* refactor render-test to use slang-rhi and remove OpenGL support
* remove -vk -glsl tests
* remove gl test
* disable failing test
* allow recursive submodules in github actions
* update slang-rhi
* update slang-rhi
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Fix redundant decorations in IRParam
Closes #4922
The problem was that same decorations were added to an IRParam multiple
times while running `specializeIRForEntryPoint()`.
`cloneGlobalValueWithCodeCommon()` kept cloning decorations for the
params that were already processed.
|
|
* Support mixture of precompiled and non-precompiled modules
This changes the implementation of precompile DXIL modules to
accept combinations of modules with precompiled DXIL, ones without,
and ones with a mixture of precompiled DXIL and Slang IR.
During precompilation, module IR is analyzed to find public functions
which appear to be capable of being compiled as HLSL, and those
functions are given a HLSLExport decoration, ensuring they are emitted
as HLSL and preserved in the precompiled DXIL blob. The IR for those
functions is then tagged with a new decoration AvailableInDXIL, which
marks that their implementation is present in the embedded DXIL blob.
The DXIL blob is attached to the IR as before, inside a EmbeddedDXIL
BlobLit instruction.
The logic that determines whether or not functions should be
precompiled to DXIL is a placeholder at this point, returning true
always. A subsequent change will add selection criteria.
During module linking, the full module IR is available, as well
as the optional EmbeddedDXIL blob. The IR for functions implemented
by the blob are tagged with AvailableInDXIL in the module IR.
After linking the IR for all modules to program level IR, the IR for
the functions marked AvailableInDXIL are deleted from the linked IR,
prior to emitting HLSL and compiling linking the result.
This change also changes the point of time when the module IR is
checked for EmbeddedDXIL blobs. Instead of happening at load time
as before, it happens during immediately before final linking, meaning
that the blob does not need to be independently stored with the module
separate from the IR as was done previously.
Work on #4792
* Clean up debug prints
* Call isSimpleHLSLDataType stub
* Address feedback on precompiled dxil support
Allow for IR filtering both before and after linking.
Only mark AvailableInDXIL those functions which pass
both filtering stages. Functions are corrlated using
mangled function names.
Rather than delete functions entirely when linking with
libraries that include precompiled DXIL, instead convert
the IR function definitions to declarations by gutting
them, removing child blocks.
* Use artifact metadata and name list instead of linkedir hack
* Use String instead of UnownedStringSlice
* Update tests
* Renaming
* Minor edits
* Don't fully remove functions post-link
* Unexport before collecting metadata
|
|
* Make sure `NullDifferential` and its witness are removed after autodiff.
* Fix.
* Add a test.
|
|
A correct semantic name is SV_DispatchThreadID with "ID" not "Index".
Those tests don't actually run and they haven't caused any problems
yet.
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
|
|
|
|
* Ignoring construct field warnings on delegatory methods
* Generalizing instruction usage type interface
* Skip collection when searching for stores
* Adding separate construct delegation tests
* Treating differentiable functions as stores
|
|
* Metal: mesh shading skeleton
* Metal: fixing mesh payload
* Metal: improving mesh shader indices output
* Metal: Implementing conditional mesh output set
* Metal: Trying to not break other backends
* Metal: trying to fix mesh output set
* Metal: Fixing MeshOutputSet usages
* Metal: Fixing vertex and primitive semantics
* Metal: Fixing code style
* Metal: Fixed hlsl indices set
* Fixed HLSL mesh output set disappearing and GLSL mesh output crashing
* Metal: Adjusting task test matching
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Allow capabilities to be used with `[shader("...")]`
Fixes: #4917
Changes:
1. Allow using capabilities instead of `Stage`s with `EntryPointAttribute`.
2. When resolving capabilities for an entrypoint+profile (per entrypoint) in `resolveStageOfProfileWithEntryPoint` add our `EntryPointAttribute` and resolved capability
3. Added tests and some capabilities related clean-up
* fix a warning made by a mistake in syntax
* change fineStageByName to assume it is passed a stage without a '_'
* test with and without prefix '_'
* cleanup some profiles and reprisentation to work better with 'Stage' and 'Profile'
This use case is why we need to clean all profile-usage into `CapabilityName`s directly.
* change how we compare
* only change profiles
* let all capabilities be resolved by 'shader' profile for now
* fix warning checks I accidently broke
* meshshading_internal to _meshshading
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
|
|
|
|
side effects (#4707)
* Adds a warning for using prefer-recompute on methods that contain side effects
* Rename `SideEffects` -> `SideEffectBehavior`
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
differentiable args (#4901)
* Explicitly detach derivative when forming a non-differentiable struct out of differentiable args
This fixes an issue where initializer lists get optimized out and lose information about non-differentiability.
There are 2 places where this could have been fixed:
1. When coercing initializer-list exprs, we can check for non-differentiable aggregate types and use a detach derivative on all the args.
2. Add an extra case in the peephole optimization step that adds detach-derivative when simplifying a make-struct of a non-differentiable type.
Even though solution 2 is more elegant, this PR goes with solution 1 simply to avoid having to use a differentiable-type-conformance-context that is used in the auto-diff IR passes to check for differentiability.
* Change test name + add expected vals
|
|
* Fix Varying Variable Location Assignments With Hull Shaders
Fixes: #4913
Fixes: #4540
Changes:
1. Added `kIROp_ControlBarrier` to HLSL/GLSL emitting.
2. Added a method to track 'used' and 'unused' varyings for when legalizing GLSL. This allows us to assign correct offsets to automatically added varyings
* Added a `ZeroLSB` check to UIntSet for this purpose
* add missing return
* code comment adjustment
* cleanup
* comment and HLSL controlBarrier mistake
* assume space for glsl/spriv varying is irrelevant
|
|
Fixes: #4899
Fixes invalid test results since `{...}` was differentiating the constructor of `myvector` when it should not (see #4877). This change modifies the test so it is correct so other PRs may be merged if indirectly/directly fixing the old use-case for this test.
|
|
* Implement `-fvk-use-dx-layout`
Fixes: #4126
Changes:
* Added fvk-use-dx-layout
* Modified `HLSLConstantBufferLayoutRulesImpl` for correctness (ex: Array is always 16 byte aligned)
* Added kFXCShaderResourceLayoutRulesFamilyImpl and kFXCConstantBufferLayoutRulesFamilyImpl to handle fvk-use-dx-layout
* Added `ConstantBufferLayoutRules` to manage constant buffer rules
* Added `alignCompositeElementOfNonAggregate`/`alignCompositeElementOfAggregate` to handle forced alignment of composites for ConstantBuffers
* `StructuredBuffer` rules are mostly equal to `scalar` layout, not much was needed to be changed to support this behavior.
* seperate legacy constant buffer and how Slang does constant-buffer normally
* undo an addition
* remove accidental test
* Address review and fix
Address review and remove GLSL support since GLSL requires a seperate legalization (need to linearlize structs like with `legalizeMetalIR` to assign explicit offsets)
* comments
* remove aggregate and non-aggregate logic
We don't need this distinction for the logic
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Fix the slang-test bug
Since we reorganize the build directory, now the libraries are
located at different directory with executables in non-Windows
platform, we have to change the code on how to find the dll directory.
* Integrate the record/replay test into slang-unit-test
We create a unit-test-record-replay.cpp to run the converted slang
examples in child process as our tests for the record-replay layer.
* Disable the test on Apple
Due to the limitation of current examples, we temporarily disable them
on apples.
Change the ci to make this test only be run on the gpu-equipped runners,
for other runners we add a white-list file
"expected-failure-record-replay-tests.txt".
* Remove 'hello-world' example from unit test
"hello-world" doesn't use gfx abstract library, instead it uses vk directly, it's
not a preferable way. So we will drop this test, instead, we will use cpu-hello-world
example.
|
|
|
|
* Fix a false failure when using the latest SPIR-V
---------
Co-authored-by: Ellie Hermaszewska <github@sub.monoid.al>
Co-authored-by: ArielG-NV <159081215+ArielG-NV@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* Support dependent generic constraints.
* Fix warning.
* Update comment.
* Fix.
* Add a test case to verify fix of #3804.
* Address review.
|
|
* Fix `tests\autodiff\reverse-while-loop-3.slang` test
Fixes: #4876
Fix `tests\autodiff\reverse-while-loop-3.slang` test to have correct listed output values
* set default to 1 (different val)
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Tuple swizzling and element access.
* Update proposal status.
* Cleanup.
* Fix merrge error.
* Address review.
|
|
* Variadic Generics Part 2: IR lowering and specialization.
* Update design doc status.
* Update design doc.
* Resolve review comments.
|
|
|
|
|
|
Previously, the warnings were:
```
environment.slang(22): warning 41022: inout parameter 'seed' is never written to
public float3 environment_sample(StructuredBuffer <Environment_sample_data> sample_buffer, inout int seed)
^~~~~~~~~~~~~~~~~~
hit.slang(5): warning 41022: inout parameter 'seed' is never written to
float3 sample_lights(inout uint seed)
^~~~~~~~~~~~~
```
With this commit they should not be emitted.
|
|
* Refactoring benchmark CI pipeline
* Combine setup and run steps
|
|
* Do not zero-initialize groupshared and rayquery variables
Fixes: #4824
`-zero-initialize` option will explicitly not:
1. Set any groupshared values to defaults
2. Set any rayQuery object to a default state (currently invalid code generation)
* grammer
* disallow groupshared initializers
disallow groupshared initializers & adjust tests accordingly
* remove disallowed groupshared-init expression
* do not default init if non-copyable
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Fix the issue that NonUniformResourceIndex is ignored
Fix the issue that after `specializeFunctionCalls`,
`NonUniformResourceIndex` is ignored in the generated specialized
function.
The reason is that if the function has a non-uniform resource parameter,
we will legalize it by replacing the resource parameter with a index,
and indexing of the resource will be moved inside the specialized function.
e.g.
```
void func(ResourceType resource) { ... }
func(resource[NonUniformResourceIndex(0)])
```
will be specialized into
```
void func(int index) { resource[index]; }
func(0);
```
In this case, inside the function, we will loose the information about
whether the resource is a non-uniform. So we add the handling for this
corner case by adding insert a `NonUniformResourceIndex` into the
specialized function:
```
void func(int index) {
int nonUniformIdx = NonUniformResourceIndex(index);
resource[nonUniformIdx];
}
```
* Fix the issue that arguments mismatch after specilization callsite
specializeCall() call could cause arguments mismatch with the parameters
of the specialized function.
For example, if the function parameter contains a resource type
```
void func(ResourceType res) { ... }
int index = ...
func(resources[index]);
```
This will be specialized into
```
void func(int index) { resources[index] }
int index = ...
func(index);
```
However, if we have more than 1 call sites, and the other call site
doesn't use `int` as the index, e.g.
```
uint index = ...
func(resources[index]);
```
this call site will be specialized into
```
uint index = ...
func(index);
```
this will be invalid, because the argument doesn't match the parameter.
so we just add the data type of the new arguments into the function key such that
For the uniformity info, we add a new attribute "IROp_NonUniformAttr",
so we will form a IRAttributedType that encodes both uniformity and data
type, and use it as the key of call info. So if there is call site using the different
data type for the resource index, we will specialize a new function for this.
* Handle the intCast and uintCast operation
Since after intCast/uintCast of nonuniformIndex, it's still a
nonuniformIndex. So we will handle this case as well.
Also, add a new test to cover this.
|
|
Adds a new Github CI action for benchmarking the slangc compiler on the MDL shaders. For now, the results are only dumped to the output of the CI, which can be later viewed through raw logs. The next step is to use github-action-benchmark to push these results into a page which will show the benchmark results over time as commits are pushed.
|
|
Fixes #4818
Emit memory qualifier once for GLSL targets.
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Support unicode identifier names.
* Fix.
* Fix language server.
* Fix build errors.
* Fix.
* Fix offset translation in language server.
|
|
Addresses #4698 as one approach to diagnose the potential problem.
Emit warnings when a user marks a parameter as `inout` but never writes to it in the function. A new intrinsic function `unmodified(out T)` has been added to explicitly indicate that an `inout` variable will not be modified in the function.
This is only one way to address the specific validation error in #4698. In general it seems that DXC does some more extensive checks on actual struct fields (as opposed to observing arbitrary struct writes), so that will be the next step.
|
|
* Fix auto-diff synthesized method naming conventions
* Update tests; remove unused var
|
|
Closes #4746.
This commit adds a support for "NumThreads" and "Shader" attribute keyword, which is in CamelCasing starting with an upper case letter.
The attribute keywords in HLSL are case-insensitive. As an example, one of D3D documents says,
"The attribute name "Shader" is case insensitive."
https://microsoft.github.io/DirectX-Specs/d3d/WorkGraphs.html
Slang, however, doesn't support the case-insensitivity. They should be all lower-case or CamelCasing starting with an upper case.
|
|
* Add embedded precompiled binary IR ops
Add IR operations to embed precompiled DXIL or SPIR-V blobs
into IR. Adds a BlobLit literal that is mostly identical to
StringLit except for its inability to be displayed, e.g.
in dumped IR. In the future, the blob might be dumped as
hexadecimal, but for now it is summarized as "<binary blob>".
* EmbeddedDXIL and SPIR-V options
The options, '-embed-dxil' and '-embed-spirv' in slangc, will
cause a target dxil or spirv to be compiled and stored in the
translation unit IR when written to a slang-module. Subsequent
changes actually implement the options.
* Per-translation unit DXIL precompilation
When -embed-dxil is specified, perform a precompilation to DXIL of
each TU, linked only with stdlib. Embed the resulting DXIL for
the TU in a IR op. Being part of IR, the precompiled DXIL can be
serialized to disk in a slang-module.
Upon loading slang-modules, the new IR op will be searched for and
the precompiled DXIL blob is saved with the loaded Module. During
linking, if all the Modules have precompiled blobs they will be
sent to the downstream compile commands as libraries instead of
source, skipping the downstream compilation, using DXC only for
linking.
Fixes Issue #4580
* Remove placeholder embedded SPIRV support
Code was added only to sketch out how other precompiled bins
will be supported.
* Remove the rest of the SPIRV placeholder support
* Fix warnings, test error on non-windows
* Remove lib_6_6 hack, add dxil_lib capability
* Allocate blob value from irmodule memarena
* Add null check after memarena allocation
* Restore the request->e2erequest code path for generatewholeprogram
* Update capability handling, move EmbedDXIL enum to end to preserve abi
* Remove lib_6_6 hack
* Move ICompileRequest functions to end
|
|
|
|
|
|
* Allow impliocit 'uniform' entrypoint parameters.
* Fix.
* Fix.
* Fix.
* Fix.
|
|
* Allow generic type deduction from ParameterBlock arguments.
* Fix test.
* Update expected failure list.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Proper warning generation for target switches and intrinsic asm
* Relaxing terminators
* Fix compiler warnings
* Rectified target switch reachability check
* Simplify target switch reachability check
* Refactoring variable names
* Using getBlocks
* Moving ad hoc special case to diagnostics source
* Using the LINE directive for testing
* Simplifying reliance on target switches
* Skipping IR generation for empty target switches
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Fix IR lowering for generic interface types.
* Fix.
* Fix.
|
|
|
|
|
|
(#4751)
|
