| Age | Commit message (Collapse) | Author |
|---|
|
* Add `IOpaqueHandle::descriptorAccess`.
* Update doc.
* fix.
|
|
* Initial support for immutable lambda expressions.
* More diagnostics, and langauge server fix.
* Language server fix.
* Fix bug identified in review.
* Add expected result.
* Update expected result.
|
|
* cuda: Improve entry handling for SV_DispatchThreadID
Fixes #6780
This commit improves CUDA entry point handling by extracting appropriate
components from DispatchThreadID based on parameter type. It now
properly handles uint scalar (x component only) and uint2 vector (x,y
components) instead of always using the full uint3 value.
Add a new test case to check for this.
* format code
* fix CI failure
* Handle review comments
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* cuda: Add support for subscript operator
This CL adds support for the subscript operator for Read Only
textures in cuda. Also adds a test for this.
Fixes #6781
* format code
* fix review comments
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* Support `this` use of `IRMeshOutputRef`
* Fix SPIR-V val error
There was a type mismatch causing a spir-v failiure: `store(var, var)` instead of `store(var, load(var))`
|
|
* Add Slang Byte Code generation and interpreter.
* Fix compile issues.
* format code
* More compile fix.
* Fix clang issue.
* Fix more clang issues.
* Another clang fix.
* Fix clang issues.
* Fix another clang issue.
* Fix wasm build.
* Update building.md
* Fix test-server.
* Fix compile error.
* Fix bug.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
|
|
generic expressions) (#6787)
|
|
* Update spirv-tools to for SDK v2025.2
Fixes: #6850
* bump spirv version to 1.4 for op linkage
* skip-spirv-validation for coop mat
* add skip-spirv-validation option to slang session desc
* use SPV_ENV_UNIVERSAL_1_6 for spirv-tool env target
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
---------
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Implemented #pragma warning
Based on https://learn.microsoft.com/en-us/cpp/preprocessor/warning?view=msvc-170
* Make #pragma warning work with #includes.
- SourceLoc are not sorted by inclusion order.
- Construct a mapping from SourceLoc to "absolute locations" that are sorted by inclusion order (roughly represents a location in a raw file with all #include resolved).
- The absolute location can be used in the pragma warning timeline
* Added preprocessor #pragma warning tests.
- Fixed #pragma warning (push / pop) SourceLoc
- Fixed unused directiveLoc in #pragma warning parsing
* #pragma warning: Added some comments and fixed some typos
* Cleaned #pragma warning preprocessor implementation.
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
Work on #6892
The function `tryInferLoopMaxIterations` in `slang-check-stmt.cpp` was
doing:
auto litExpr = m_astBuilder->create<LiteralExpr>();
but the declaration of `LiteralExpr` in `slang-ast-expr.h` had been
marked as abstract, during the switch to using the `slang-fiddle` tool
to generate code:
FIDDLE(abstract)
class LiteralExpr : public Expr
{ ... }
In this case, the intention of the AST design is that `LiteralExpr`
should be kept abstract, and only the concrete subclasses should
ever be instantiated.
Because of some historical design choices, the `ASTNodeType`
enumeration includes both the concrete and abstract AST classes, so
the code ended up constructing a `LiteralExpr` that had the tag
`ASTNodeType::LiteralExpr`. Then attempts to use the `ASTNodeDispatcher`
code on such a node caused crashes, because the dispatcher code only
included `case` statements for the non-abstract `ASTNodeType`s.
The quick fix here was to change the `tryInferLoopMaxIterations`
function to instead do:
auto litExpr = m_astBuilder->create<IntegerLiteralExpr>();
A test case was added to help catch any future regressions on this
specific issue.
A more long-term fix should involve introducing code that statically
and/or dynamically prohibits the creation of instances of AST node
classes that have been marked abstract.
|
|
* update slang-rhi submodule
* slang-rhi API changes
* disable agility sdk
* fix texture creation
* update formats in tests
* Extent3D rename
* use 1 mip level for 1D textures for Metal
* fix texture upload
* update to latest slang-rhi
* update slang-rhi
* format code
* update slang-rhi
* do not run texture-intrinsics test on metal
* update slang-rhi
* deal with failing tests
* fix more tests
* update slang-rhi
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Simon Kallweit <simon.kallweit@gmail.com>
|
|
* Fix #6544: Properly format nested type names in extensions
Modify DeclRefBase::toText to properly handle types defined
in extensions by qualifying them with their parent type name.
This ensures getFullName() returns the full name like
'FullPrecisionOptimizer<half>.State' instead of just '.State'.
Also handle other nested types in structs/classes similarly.
* Update extension reflection handling - with generics args and namespaces
- stopping namespace inclusion for extension members
- Update to use getTargetType() to handle the generic arguments
- update test cases
* Simplify code to remove using parentDecl
|
|
* Initial implementation for SPIRV, GLSL and Metal
* test add bool test
* Fix and improve subgroup rotate tests
* Add proper GLSL extensions and proper Metal type checking
* Clean up tests and add diagnostics test for subgroup type for Metal
* Update wave-intrinsics docs
|
|
* Add struct member offset qualifier for SPIRV
* Implement for GLSL target and add tests
* clean up
* fix formatting
* fix typo
* renamed GLSLStructOffset to VkStructOffset and added emit-spirv-via-glsl test case
|
|
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
|
|
Change modifies the countbits intrinsic to use generics in order to
support 64bit countbits on select platforms where this is supported.
On platforms where this is not natively supported, we emulate by
converting the 64-bit type into a uint2 (metal and spir-v).
This should align with the implementation of other uint64_t
intrinsics such as abs, min, max and clamp.
Added new countbits64 test to verify changes.
Updated documentation for 64bit-type-support.html
|
|
Close #6589.
In PR #6487, we support partial specialization. However there is a corner case we didn't handle correctly.
For the IR like this:
%val: specialize(...) = some inst;
%arg1: specialize(...) = makeExistential(%val, ...);
%arg2: %SomeConcreteType: load(...);
call func(%arg1, %arg2);
when we specialize the call func instruct, we will also specialize the function parameters.
On our existing logic, when we find an argument is a makeExistential, we will always extract the existential value, and use its type as the new parameter.
But in this case, %arg1 is not fully specialized yet, so it's type will still be a specialize. In this case, we will change the function's first parameter from an existential type to a specialize. This will result in that we lose the chance to specialize the first argument in the next iteration, because the first parameter of this function is not an existential type any more.
The reason behind this is that we should always keep specializing the arguments and parameters at the same time.
So this PR just does a check before specializing the parameters that if the argument cannot be fully specialized, we won't specialize the parameter this early. Instead, we will wait for the next iteration until the argument can be specialized.
|
|
|
|
(#6841)
With this PR, cooperative-vector tests will run only when the hardware can support the feature.
Unfortunately, DX12 backend will always reports that cooperative-vector is NOT supported until it is officially supported in May.
|
|
* initial wip for spirv
* working tiled example
* clean up store and load
* minor fixes
* fix loadAny name
* add initial tests, including broken/unimplemented intrinsics
* fix subscript
* run tests at 16x16, remove not supported arithmetic tests
* minor fixups on implementation
* rename CoopMatMatrixUse
* Update tests to pass validation layers locally
* Add mat-mul-add test and minor fixes
* Add more tests
* Remove dead code
* Add coopMatLoad function and tests, enforce constexpr for matrix layout
* Use getVectorOrCoopMatrixElementType in place of getVectorElementType
|
|
* Fix matrix division by scalar for Metal and WGSL targets
* Add tests
* Minor fix
* Fix compilation error
* Convert to multiplication for WGSL
* Minor cleanup
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Fix User Attribute string reflection
Fixes #6794
* Fix strings not being properly escaped
---------
Co-authored-by: Darren Wihandi <65404740+fairywreath@users.noreply.github.com>
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Add gl_PointCoord support in GLSL compat mode
* Add SV_PointCoord
* Test on metal as well
* Update SPIRV system value semantics table in docs
* Update metal docs for SV_PointCoord
|
|
* Fix pointer field access for GLSL
* Add test
* Fix SPIRV test
* add spirv via glsl test
|
|
Fixes #6171
This commit adds logic for reporting double support
to the d3d11 backend, for running tests on GPUs that
do not support D3D11_FEATURE_DOUBLES, and add checks
for that support to tests that require the feature.
|
|
This fixes issue #6654
Only hoist instructions that are optimized by prepareFuncForForwardDiff.
Add flag hoistLoopInvariantInsts to IRSimplificationOptions and set this
to true only if called from prepareFuncForForwardDiff, then only hoist
if the flag is set. Additionally, do not hoist loops if they only have a
single trivial iteration.
|
|
interface-typed output parameter (#6788)
* More specific diagnostic for invalid concrete-to-interface arg coercion
* Add test for the new error message
* Fix typo in expected test result
|
|
* Get real value for typeAdapter
When the type is mismatch and typeAdapter is used, get the real value
from typeAdapter so that we don't get nullptr for irValue.
This fixes the assert if uint is used for SV_VertexID, which is an int
in the system binding semantic.
Fixes: #6525
* Add test case; add nullptr check
|
|
Closes https://github.com/shader-slang/slang/issues/5995
|
|
* Add support for Ray Payload Access Qualifiers (PAQs) (#3448)
- Added [raypayload] attribute for struct declarations
- Implemented field validation requiring read/write access qualifiers
- Added diagnostic error for missing qualifiers
- Enabled PAQs in DXC compiler and HLSL emission
- Added new test demonstrating PAQ syntax
- Implemented proper handling of ray payload attributes in IR generation
* format code
* Cleanup: Remove unused vars
* Add check to enablePAQ only for profile >= lib_6_7
* Review Fix - Add PAQ support for DX Raytracing
add enablePAQ flag to DownstreamCompileOpitons, improve PAQ handling
update raypayload-attribute-paq.slang to ensure hlsl and dxil is
validated
* Add diagnostic test for missing paq for lib_6_7
Compile using `-disable-payload-qualifiers` aka lib_6_6 profile
raypayload-attribute-no-struct.slang and
raypayload-attribute.slang
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
|
|
(#6651)
* Fix incorrect assert on mixed global uniform and varyings
* add test
* remove unnecessary include
* fix incorrect logic
* fix comment grammar
* address review comments and improve test
* minimize diff
* fix more issues for cuda build
* remove unnecessary line for diff
|
|
|
|
(#6696)
* Initial loop analysis pass
* More changes for a single-pass implication propagation
* Update slang-ir-autodiff-loop-analysis.cpp
* Cleanup + new system for loop analysis
* Fixup bugs in loop analysis
* Remove some relation types to simplify the analysis. Add test
* Remove unused
* Address comments
* Fix issue with continue loops
* Update reverse-loop-exit-value-inference-1.slang
* Update reverse-continue-loop.slang
|
|
output) function outputs (#6737)
Closes https://github.com/shader-slang/slang/issues/6632
|
|
* fix(d3d11): correct parameter in VSSetConstantBuffers1 from uavCount to cbvCount (fixes #6531)
Root cause - Incorrect parameter passing in slang-rhi
1. slang-rhi #281 - Add the correct cbvCount for setting Constant Buffer
2. Prevent render tests from overwriting reference images
* Add missing tests/render/multiple-stage-io-locations.slang.3.expected.png
* Add more expected images from texture2d-gather
* Add new option: skipReferenceImageGeneration
For Github CI we set this to true - So we don't overwrite the expected
images
---------
Co-authored-by: Jay Kwak <82421531+jkwak-work@users.noreply.github.com>
|
|
* Implement sparse texture Load intrinsics for SPIRV
* changed test name from TEST_load to TEST_sparse
---------
Co-authored-by: Darren Wihandi <65404740+fairywreath@users.noreply.github.com>
|
|
* Fixed generic interface specialization crashes:
- Add an export decoration to specialized generic interfaces.
* Fixed generic interface specialization crashes:
- Add an export decoration to specialized generic interfaces.
- Use getTypeNameHint(...) instead of a manual mangler.
* In cloneInstDecorationsAndChildren: specialize all linkage decorations, not just the exports.
- If a linkage decoration is already present, it is not specialized and replaced by the specialized one.
- If a specialization uses the TypeNameHint, sanitize it to be used as an identifier.
- Use the identifier name sanitizer from slang-mangle.
* Added tests/generics/generic-interface-linkage.slang
- See #6601 and #6688
|
|
This PR enables existing CoopVec tests with DX12 backend.
In order to use the CoopVec feature with DX12 backend, we have to use an option, "-dx12-experimental", because the current implementation of CoopVec feature in dxcompiler.dll requires "experimental feature".
Note that when the "experimental feature" is enabled, slang-test becomes less stable.
For that reason, we should use the option "-dx12-experimental" only when it is needed.
All tests for GLSL are deleted because CoopVec support for GLSL in Slang is deprecated and no longer supported.
Some of CoopVec tests are still disabled for DX12 backend because:
DXC doesn't support 8bit integer types and
Some of CoopVec features are not implemented in DXC backend.
|
|
* Reapply "Eliminate empty struct on metal target (#6603)" (#6711)
This reverts commit bc9dc6557fc0cc3a4c0c2ff27e636940e361cf5d.
* Remove argument in make_struct call corresponding to void field
This is a follow-up of #6543, where we leave the VoidType field as it in
make_struct call during legalization pass.
So during cleaning_void IR pass, when we remove "VoidType" from struct,
we will have to also clean up the argument corresponding to the
"VoidType" field.
|
|
Fixes issue #6533
This patch updates handling of Array and ConstantBuffer types for WGSL
transpiling, giving correct syntax for arrays of buffers in WGSL.
|
|
* Use coopvec supporting dxcompiler.dll and dxil.dll
* Fix the failing tests
|
|
* Add GetDimensions support for CUDA
This CL adds GetDimensions support for cuda by using the PTX
instructions. Currently, PTX only supports getting width, height and
depth.
This CL also adds a new test to test this support.
Fixes #5139
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
# Make `IRWitnessTable` Hoistable
## Intention of the PR
This commit makes `IRWitnessTable` Hoistable so that we can avoid duplicated `IRWitnessTable`.
## Problems
This commit tries to address the following issues arise after turning `IRWitnessTable` into Hoistable:
1. A Hoistable instance is immutable.
2. When tries to create a duplicated child, you will get a previously created instance of `IRWitnessTable`, instead of a new one.
3. We don't actually want to hoist `IRWitnessTable`.
4. There can be only one instance of Hoistable and it cannot appear as childs multiple times.
5. Different import/export mangled names were used for the same Witness-table when its type is "enum" interface.
## Implementation
### Solution for "1. A Hoistable instance is immutable."
`IRWitnessTable::setConcreteType()` is removed, because when an `IRInst` is Hoistable, it is treated as immutable. Any `IRInst::setXXX()` methods don't work anymore.
There were two places calling `setConcreteType()` and their logic had to change little bit.
`DeclLoweringVisitor::visitInheritanceDecl()` in `source/slang/slang-lower-to-ir.cpp` was calling `setConcreteType()`. It had a little strange logic around `lowerType()`. The `IRWitnessTable` was added with `context->setGlobalValue()` first and its `concreteType` was changed later. This commit works around in a way that it sets the parent of `IRWitnessTable` temporarily and reset it with the correct `IRWitnessTable`. Without this logic, it went into an infinite recursion.
`AutoDiffPass::fillDifferentialTypeImplementation()` in `source/slang/slang-ir-autodiff.cpp` was calling `setConcreteType()`. It was changing the concreteType of `innerResult.diffWitness`. This commit creates a new `IRWitnessTable` and copies its `IRWitnessTableEntry`.
### Solution for "2. When tries to create a duplicated child, you will get a previously created instance of IRWitnessTable, instead of a new one"
After a call to `IRBuilder::createWitnessTable()`, this commit checks if the returned `IRWitnessTable` is a brand new or not. If it is not a new one, we have to avoid adding the decorations and children.
This commit decides when to add decorations and children based on whether `IRWitnessTable` has any of decorations or children already. It doesn't seem like a proper way to check. But when I tried, it was difficult to find a bottleneck point where the decorations and children are added to `IRWitnessTable` first time. Note that we are not trying to find when `IRWitnessTable` is created for the first time; we need to find if the decorations and children were added once.
It might be fine to have duplicated `IRWitnessTableEntry` in most of the cases, but I noticed that it fails an assertion check when `shouldDeepCloneWitnessTable()` returns false in `cloneWitnessTableImpl()`.
### Solution for "3. We don't actually want to hoist IRWitnessTable."
The reason why this commit makes `IRWitnessTable` is to prevent the duplicated instances of `IRInst`. But we don't really want to "Hoist" them.
When an `IRWitnessTable` gets Hoisted out, it causes unexpected problems and the specialization process fails due to the missing `IRWitnessTable` in the input.
This commit prevent from hoisting `IRWitnessTable` in `_replaceInstUsesWith()`. The way this is implemented feel little hack but we discussed on Slack and decided to go with this. One of the proper approaches could be to add a new flag in `IROpFlags` and have a new one like `kIROpFlag_Deduplicate`, which is different from just `kIROpFlag_Hoistable`.
### Solution for "4. There can be only one instance of Hoistable and it cannot appear as childs multiple times."
When `IRWitnessTable` is Hoistable, there can be only a unique set of instances. And we cannot have an instance as a duplicated childs. It is because `IRInst` has only one set of `IRInst* next` and `IRInst* prev`.
Before this commit, an instance of `IRGeneral` could have duplicated instances of `IRWitnessTable`. As an example, `IInteger` interface inherits two other interfaces, `IArithmetic` and `ILogical`. And they both inherits from `IComparable`.
```
interface IInteger : IArithmetic, ILogical {}
interface IArithmetic : IComparable {}
interface ILogical : IComparable
```
When we specialize it in `specializeGenericImpl()`, an `IRBlock` gets the following list of children:
- IRWitnessTable for IComparable,
- IRWitnessTable for IArithmetic,
- IRWitnessTable for IComparable,
- IRWitnessTable for ILogical,
For the cloning during the specialize, "IRWitnessTable for `IComparable`" must be cloned before the cloning of "IRWitnessTable for `IArithmetic`". Because "IRWitnessTable for `IArithmetic`" refers "IRWitnessTable for `IComparable`" as its `IRWitnessTableEntry`. The order they appear in the `IRBlock` as children decides which instances will be cloned first. And "IRWitnessTable for `IComparable`" must appear before "IRWitnessTable for `IArithmetic`".
Note that "IRWitnessTable for `IComparable`" appears twice, The first one was added for "IRWitnessTable for `IArithmetic`". And the second one is added for "IRWitnessTable for `ILogical`".
With this commit "IRWitnessTable for `IComparable`" can appear as a child only once in `IRBlock`. So it causes an error if it gets the following list:
- IRWitnessTable for IArithmetic,
- IRWitnessTable for IComparable,
- IRWitnessTable for ILogical,
In order to resolve the problem, "IRWitnessTable for `IComparable`" must appear before both "IRWitnessTable for `IArithmetic`" and "IRWitnessTable for `ILogical`" as following:
- IRWitnessTable for IComparable,
- IRWitnessTable for IArithmetic,
- IRWitnessTable for ILogical,
To address the problem, the instances of `IRWitnessTable` is always added to the end of the children list. If it is already added to the list, we don't move. This works out because the AST tree is built based on the dependencies.
### Solution for "5. Different import/export mangled names were used for the same Witness-table when its type is "enum" interface."
This issue was found while testing with Falcor tests where it uses Conformance-type feature of Slang.
We are using different import and export mangled names for a same Witness-table when the witness-table is for "Enum" interface.
The way we simplify the implementation of "Enum" causes a problem when it comes to generate export/import for the witness-table. And the exact repro step is still unclear.
There were two suggested solutions for the problem and this PR adopted the first option for now. Maybe we want to improve it with the second option later.
option 1, when we produce mangled names for those witness-table, we can use a mangled name with the underlying "int" type instead of the name of the enum type. In this way, all witness-tables for enum types whose underlying type is same will get the same mangled name. It will allow us to deduplicate the witness-table during the linking.
option 2, we can preserve type info for enum type when generating IR. We can still erase all other uses of the type info of enum types for now. But when we generate the witness-table, instead of filling the conforming type operand to IntType, we fill it as EnumType(IntType) where EnumType is a new global IROp code to represent all enum types (like InterfaceType/StructType). This way the operands for the two witness-tables will be different.
"option 1" is more quick and dirty and "option 2" is more proper way to address it.
I should go with "option 1" and improve it with "option 2" approach later.
|
|
* Include generics' operands in call graph construction
* add test
|
|
This reverts commit b3deec2001ea34e20e9a6af8ddf5cf3866cafac0.
|
|
* Eliminate empty struct on metal target
Close 6573.
We previously disabled the type legalization for ParameterBlock on
Metal, but Metal doesn't allow empty struct in the argument buffer
which is mapped from ParameterBlock, so we will need legalizeEmptyTypes
on Metal target.
* update test
* update function name
|
|
* Fix SPV_KHR_maximal_reconvergence extension name spelling
Vulkan validation layers emit warnings on lowercase khr.
* Move OpExtension check
|
