| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
* 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
|
|
* Document matrix multiplication operators
This is to address one of cases where a user was confused about why
Slang doesn't have operator overloading of multiplication for matrices.
It is little confusing that glsl.meta.slang implemented `operator*()`
for matrices but Slang/HLSL doesn't. It turned out that HLSL/Slang
performs a component-wise multiplication with `operator*()` whereas GLSL
performs a matrix multiplication.
* Mentioning CUDA
* Update based on the review feedback
|
|
This PR fixes the compiler warning like following,
lld-link: warning: ignoring unknown argument '--no-undefined'
lld-link: warning: ignoring unknown argument '--build-id', did you mean '-build-id'
chatGPT said that those options are for ELF and not for Windows:
These warnings happen because the Clang toolchain on Windows is invoking lld-link, which is the LLVM linker frontend for link.exe compatibility (i.e., it acts like the MSVC linker). The arguments --no-undefined and --build-id are ELF-specific flags, which are valid when targeting Linux, but not valid in the PE/COFF (Windows) environment.
Here's a breakdown:
⚠️ Warning messages explained:
lld-link: warning: ignoring unknown argument '--no-undefined'
→ This is a Linux/ELF flag. On Windows, this has no effect and is unknown.
lld-link: warning: ignoring unknown argument '--build-id', did you mean '-build-id'
→ Again, --build-id is for ELF binaries to include a unique build ID. Not valid for Windows targets.
|
|
* Check the available VK extensions before using CoopVec APIs in GFX
* Remove a redundant request for cooperative vector extension for vk
|
|
* Added Dictionary::erase(iterator) and fixed crashing when filtering a dictionary in slang-ir-autodiff-loop-analysis.cpp
* Added Dictionary::removeIf(Predicate)
* Removed Dictionary::erase(it)
---------
Co-authored-by: Julius Ikkala <julius.ikkala@gmail.com>
|
|
|
|
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.
|
|
* Add Yet Another Source Code Generator
This change introduces an offline source code generation tool,
provisionally called `fiddle`. More information about the design of
the tool can be found in `tools/slang-fiddle/README.md`.
Yes... this is yet another code generator in a project that already
has too many. Yes, this could easily be a very obvious instnace of
[XKCD 927](https://xkcd.com/927/).
This change is part of a larger effort to change how the AST
types are being serialized, and the way code generation for them
is implemented.
Right now, the source code for the new tool is being checked in and
the relevant build step is enabled, just to make sure everything is
working as intended, but please note that this change does *not*
introduce any code in the repository that actually makes use of
the new generator. All of the AST-related reflection information that
feeds the current serialization system is still being generated using
`slang-cpp-extractor`.
The design of the new tool is primarily motivated by the new approach
to serialization that I'm implementing, and once that new approach
lands we should be able to deprecate the `slang-cpp-extractor`.
In addition, the new tool should in principle be able to handle
many of the kinds of code generation tasks that are currently being
implemented with other tools like `slang-generate` (used for the core
and glsl libraries). This tool should also be well suited to the task
of generating more of the code related to the IR instructions.
* format code
* Build fixes caught by CI
* Fix another warning coming from CI
* Another CI-caught fix
* Change bare hrows over to more proper abort execptions
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Eliminate back-reference in ChildStmt
This change is part of a larger effort to improve the code for AST
serialization in the Slang compiler.
Tree structures are understandably easier to serialize than DAGs,
and DAGs are easier than fully generaal graphs.
The Slang AST nodes form a tree structure... except when they don't.
Among the exceptions to nice tree-structured ASTs are:
1. References to `Decl`s are encoded as pointers to the AST `Decl`
nodes themselves. This can result in cycles in the graph, and
requires care in serialization.
2. Nodes that inherit from `Val` represent, well, *values* instead
of actual pieces of syntax, and as such they are deduplicated so
that identical values will (hopefully) be identical pointers.
This results in a DAG structure for `Val`s, but at least it's not
a general graph (except for cycles that go through a `Decl`).
3. There are some minor cases of DAG-structured sharing that the
parser can introduce to deal with cases when a traditional-style
declaration includes multiple declarators. E.g., given:
```
static int a, b;
```
The resulting `DeclGroup` will include distinct `Decl`s for `a`
and `b`, which will share the `static` modifier through a
`SharedModifiers` node, and the `int` type specifier through a
`SharedTypeExpr` node.
This duplication can be ignored, for the purposes of serialization,
since duplicating those parts of the AST has no major down-sides.
4. There is the case of `ChildStmt`, used for things like `break`
and `continue`, which stores a direct `Stmt*` to the enclosing
parent statement being targetted. Storing the target is useful so
that IR lowering doesn't need to repeat the work that the semantic
checking logic did to associate each child statement with its parent.
The parent link inside of `ChildStmt` creates a cycle in the AST
`Stmt` hierarchy, since the outer statement contains the inner,
and the inner statement stores a pointer to the outer.
This change eliminates the last of these sources of complication for
AST serialization, by changing the `ChildStmt` type to stored an
integer ID for the enclosing statement that it matches to, and having
each `BreakableStmt` (used to represent the outer `switch`, or loop,
or whatever) generate its own unique ID as part of semantic checking.
Note: if necessary, it is reasonable for the outer statement to have
its unique ID generated as part of parsing, rather than semantic
checking.
* format code
* Change unique ID to be a proper Decl
The fix here is to make the "unique ID" representation be a full
`Decl`-derived AST node, so that it is both allowed to break the
tree-structuring rules cleanly, and it is also trivially guaranteed
to be unique across all loaded ASTs.
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Fix compiler warning with clang 18.1.8 on windows
|
|
|
|
(#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.
|
|
* Remove support for ad hoc Slang IR compression
This change is part of a larger effort to clean up the approach to
serialization in the Slang compiler. The overall goal is to simplify
and streamline all of the serialization-related logic, so that we are
left with code that is less "clever," and easier to understand for
contributors to the codebase.
Removing support for compression of serialized Slang IR has
benefits that include:
* Reduction in code complexity: consider things like the subtle way
that the `FOURCC`s for compressed chunks were being computed from
the uncompressed versions, and the mental overhead that goes into
understanding that, for anybody who would dare to touch this code.
* Reduction in testing burden: there have been, de facto, two
very different code paths for serialization of the Slang IR, and
it is not clear that the existing test corpus for Slang has
sufficient coverage for both options. By having only a single code
path, every test that performs any amount of IR serialization helps
with test coverage of that one path.
* Opportunity to explore alternatives. This is perhaps a reiteration
of the first point, but once the code is stripped down to the
simplest thing that could possibly work (I am not claiming it has
reached that point yet), it becomes easier for contributors to
understand, and it becomes more tractable for somebody to come along
with an improved approach that performs better (in either
compression ratio or performance) while still being maintainable.
In my own local setup, I found that removing support for Slang IR
compression led to the `slang-core-module-generated.h` file increasing
in size from 46.1MB to 47.4MB. This increase in the `.h` file size
for the core library binary only resulted in a release build of
`slang.dll` increasing from 20.0MB to 20.2MB. Removing the ad hoc
compression support has almost no impact on the size of actual binary
Slang modules *so long* as the additional LZ4 compression step is
being applied to them.
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Use the latest Ubuntu version not specific old version
|
|
* 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
|
|
Documenting CoopVec related functions.
This commit also fixes a few warning printed from the doc generation tool.
Some of comments are removed or converted from /// to //, because the overloading functions can have /// style comment only once.
|
|
* Consume `;` after parsing typedef decl.
* Fix.
* Fix regressions.
|
|
The auto labeling CI has been failing as a result of the
SLANGBOT_PAT token not having all of the necessary permissions.
Change updates the tokens used for requests to use new tokens that
were created with the specific permissions needed for the operations
in this CI workflow.
|
|
* 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>
|
|
Co-authored-by: Simon Kallweit <simon.kallweit@gmail.com>
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 commit resolves three issues in the benchmark script:
1. Fixed nested f-string syntax error that was causing CI failures (fixes #6772)
2. Fixed key access error handling for timing dictionary
The nested f-string was causing a syntax error in CI, and the proper fix is to
create the key variable separately before using it. Also improved error handling
for cases where compilation fails and timing data might be missing.
Fixes: #6772
|
|
* Fix user-guide typos
Use LLM to scan each of the markdown files to fix typos.
Try not to change anything but the typos in this CL.
* typo not caught by LLM
* add output of ./build_toc.ps1
|
|
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
|
|
Change introduces a workflow that will automatically run on new
issues and add a "Dev Opened" label if opened by select dev
members. This workflow can be used to add additional labels in
the future.
|
|
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* 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
|
|
* void field rework
* move void cleanup pass earlier
|
|
Closes https://github.com/shader-slang/slang/issues/5995
|
|
Fixes #6624
This commit changes the behavior of getArgumentValueString() to return
the string's value, instead of returning the string's token,
as that token also contains the surrounding quotation marks.
This commit also modifies the relevant unit test accordingly,
to not check for the surrounding quotations.
|
|
* Use GITHUB_TOKEN for fetching prebuilt
This PR extends Commit c6b702c to use GITHUB_TOKEN if set for fetching
prebuilt binaries.
This allows webgpu-dawn and slang-tint to be downloaded for certain IPs
where the github API rate is limited.
Fixes #6689
* Don't ignore download failure if github token is provided
* Update readme for getting github access token
* format code
* combine cmake_parse_arguments calls
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Jay Kwak <82421531+jkwak-work@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* 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>
|
|
* implement parameterblock for metal
Metal uses argument buffer to pass parameter buffer to pipeline, in this
change, we implement a simple way to copy the data to argument buffer.
In argument buffer tier2 rule, all the fields in parameter block will be
flatten to ordinary data, therefore
- we keep the m_data as in ShaderObjectImpl a CPU buffer to track on the data set in.
- For resource types, they will be represeted as device pointer or resource id in argument
buffer, we will just set their address or id at corresponsing offset in the CPU buffer
every time when 'setResource' or 'setSampler' is called.
- When binding the pipeline, we just simply copy the CPU argument buffer to GPU argument buffer.
- The only special case is nested parameter block. Because nested parameter block is represented
as a device pointer which will be another argument buffer, we will just recursively call
`_ensureArgumentBufferUpToDate` to get sub-object's argument buffer, and fill the GPU address of
those 'sub'-argument buffer to the root argument buffer at correct offset.
* Inform command encoder to hazard track the bindless resources
Since for all the resources within argument buffer are bindless, Metal
won't automatically hazard track those resources, we will have to call
'useResources' to inform Metal to hazard track those resources,
otherwise we will have to call wait fence after each command submission.
* nullptr check
* address comment
|
|
* 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
|
