| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This PR implements `Access.Immutable` to allow pointers to immutable
data.
The new type `ImmutablePtr<T>` is defined as an alias of `Ptr<T,
Address.Immutable>`.
By forming a immutable pointer, the programmer is conveying to the
compiler that the data at the pointer address will never change during
the execution of the current program. Therefore loads from immutable
pointers can be deduplicated by the compiler, and will translate to
`__ldg` when generating code for CUDA.
The SPIRV backend is not changed in this PR, since the current SPIRV
spec makes it very difficult to specify loads from immutable address
without generating tons of wrappers and boilerplate type declarations.
We would like to see the spec evolved a bit to around its support of
`NonWritable` physical storage pointers or immutable loads before we
attempt to express such immutability in SPIRV. For now we simply emit
ordinary pointers and loads when generating spirv.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The varying param legalization pass didn't deal with this 1D form of
SV_DispatchThreadID for CPU targets:
```slang
void computeMain(int i : SV_DispatchThreadID)
```
Instead, it just overrode the type of `i` with a `uint3`, breaking lots
of code that attempted to use `i` for something, like a `switch`
statement for example.
I ran across this when going through `language-feature` tests for the
LLVM target, which will also use this legalization pass. I'm separately
submitting this now because this also fixes the existing CPU target. The
test I enable in this PR is one that was previously generating broken
code on CPU.
(somewhat related issue: #7468)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(#8547)
This allows us to specialize functions whose argument is a sub element
of a constant buffer, instead of being only applicable to entire buffer
element. Closes #8421.
This change also implements a proper heuristic to determine when to
specialize the calls and defer the buffer loads.
This PR addresses a pathological case exposed in
`slangpy\slangpy\benchmarks\test_benchmark_tensor.py`, which used to
take 27ms to finish, and now takes 1.25ms.
For example, given:
```
struct Bottom
{
float bigArray[1024];
[mutating]
void setVal(int index, float value) { bigArray[index] = value; }
}
struct Root
{
Bottom top[2];
[mutating]
void setTopVal(int x, int y, float value)
{
top[x].setVal(y, value);
}
}
RWStructuredBuffer<Root> sb;
[shader("compute")]
[numthreads(1, 1, 1)]
void compute_main(uint3 tid: SV_DispatchThreadID)
{
sb[0].setTopVal(1, 2, 100.0f);
}
```
We are now able to specialize the call to `setTopVal` into:
```
void compute_main(uint3 tid: SV_DispatchThreadID)
{
setTopVal_specialized(0, 1, 2, 100.0f);
}
void setTopVal_specialized(int sbIdx, int x, int y, float value)
{
Bottom_setVal_specialized(sbIdx, x, y, value);
}
void Bottom_setVal_specialized(int sbIdx, int x, int y, float value)
{
sb[sbIdx].top[x].bigArray[y] = value;
}
```
And get rid of all unnecessary loads. Achieving this requires a
combination of function call specialization and buffer-load-defer pass.
The buffer-load-defer pass has been completely rewritten to be more
correct and avoid introducing redundant loads.
This PR also adds tests to make sure pointers, bindless handles, and
loads from structured buffer or constant buffers works as expected.
|
| |
|
| |
Add 29 intrinsics to the list by new test
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
logic (#8168)
Fixes: #8167
Current emitting logic does not work, this has been corrected.
The provided test ensures our CUDA code is valid by compiling PTX from
it.
`m_writer->emit("OptixTraversableHandle");` should be `out <<` since
`out` adds to type-name-cache; otherwise using a type twice will produce
bad type-names (since we filled type-name cache with "" instead of
"typeName")
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Due to an older version of spec referred there was an inconsitency v1.29
2/20/2025 - [HitObject LoadLocalRootArgumentsConstant]
Latest spec
https://microsoft.github.io/DirectX-Specs/d3d/Raytracing.html#hitobject-loadlocalroottableconstant
Refer:
OptiX backend support for Shader Execution Reordering (SER) features as
outlined in issue #6647. -
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes #7574
Changes:
* Add an initial (fairly simple) optimization pass which is able to
eliminate redundant copies.
* Our current existing optimizer passes remove redundant load/store very
robustly, this pass will focus on other cases of copy elimination
* Primary approach is to make all functions which are `in T` and `T` is
trivial to copy into a `__constref T`. We then (depending on scenario)
manually insert a variable+load if a pass-by-reference is not possible;
otherwise we pass by `constref`.
* Added optimizations to eliminate redundant code which causes
`constref` to fail to compile
---------
Co-authored-by: Harsh Aggarwal <haaggarwal@nvidia.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix 7441: CUDA boolean vector layout to use 1-byte elements
Boolean vectors (bool1, bool2, bool3, bool4) were incorrectly implemented
as integer-based types using 4 bytes per element instead of actual 1-byte
boolean elements on CUDA targets.
Changes:
- Update CUDA prelude to define boolean vectors as structs with bool fields
instead of typedef aliases to integer vectors
- Implement CUDALayoutRulesImpl::GetVectorLayout to use 1-byte alignment
for boolean vectors, matching actual CUDA memory layout behavior
- Update make_bool functions to populate struct fields correctly
This ensures boolean vectors have the same memory layout as bool[4] arrays:
- bool1: 1 byte (was 4 bytes)
- bool2: 2 bytes (was 8 bytes)
- bool3: 3 bytes (was 12 bytes)
- bool4: 4 bytes (was 16 bytes)
Fixes memory layout mismatch between Slang reflection API and actual
CUDA compilation, achieving 75% memory savings for boolean vector usage.
* Fix CI issues -
Add and update associated functions and operators
* Make boolX same as uchar
* Use align construct on struct for boolX
* Improve Test case for robust alignment checks
* Formatting
* Disable selected slangpy tests
* add metal check which is slightly different than cuda
* Test-1
* Test-2
* Test-3
* Test-4
* ReflectionChange
* cleanup and update
* _slang_select with plain bool is needed for reverse-loop-checkpoint-test
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add new capdef for lss intrinsics
Fixes #7426
Raygen shaders need to be supported for only hitobject APIs. So we need
a special capability for that, instead of a common one.
* regenerate command line reference
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* WiP: Add coopvec support for Optix
* format code
* fix minor issues
* Fix review comments
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Enable LSS hit object test
Enabled LSS SER tests now that PR #7211, which added SER support to OptiX,
has been merged.
Ran: ./build/Debug/bin/slangc.exe tests/cuda/lss-test.slang -target ptx
-Xnvrtc -I"C:/ProgramData/NVIDIA Corporation/OptiX SDK 9.0.0/include"
and confirmed that the HitObject intrinsic is called.
eg:
call (%f15, %f16, %f17, %f18, %f19, %f20, %f21, %f22),
_optix_hitobject_get_linear_curve_vertex_data, ();
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* WiP: LSS intrinsics: initial commit
* format code
* Fix CI failures
* Address review comment
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
to (#6326) (#7194)
* Fix operator precedence in OptiX ray payload pointer casting
Added extra parentheses around the cast to ensure proper operator precedence when
dereferencing the OptiX ray payload pointer. This fixes the issue where the compiler
was treating the expression as (RayPayload_0 *)getOptiXRayPayloadPtr()->color_0 instead of
((RayPayload_0 *)getOptiXRayPayloadPtr())->color_0.
Error:
nvrtc 12.9: tests/cuda/optix-cluster.slang(17): error : expression
must have pointer-to-class type but it has type "void *"
nvrtc 12.9: note : (RayPayload_0
*)getOptiXRayPayloadPtr()->color_0 = color_1;
nvrtc 12.9: note : ^
Tested using:
./build/Debug/bin/slangc -target ptx -Xnvrtc
-I"/home/haaggarwal/NVIDIA-OptiX-SDK-9.0.0-linux64-x86_64/include"
-DSLANG_CUDA_ENABLE_OPTIX -entry closestHitShaderA
./tests/cuda/optix-cluster.slang
* Fix Check
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Implement shader execution reordering support for OptiX
Added OptiX backend support for Shader Execution Reordering (SER) features as outlined in issue #6647. This implementation:
1. Added CUDA target support for HitObject API
2. Implemented core SER functionality (TraceRay, MakeHit/Miss, Invoke)
3. Added OptiX-specific hit object handling functions
4. Added test case for OptiX SER functionality
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add noreturn attribute to IgnoreHit
* Revert "Add noreturn attribute to IgnoreHit"
This reverts commit 3cf2354dada71b9a8713b08f3a2e261de4aabfa4.
* Fix: Preserve inout param modifications with OptiX IgnoreHit()
Issue #6326 identified that in OptiX, when using IgnoreHit() (which
maps to the "noreturn" optixIgnoreIntersection() intrinsic), any
modifications made to 'inout' parameters within the shader would be
lost. This was due to IgnoreHit() preventing the execution of the
copy-back operation from the temporary variable (used to implement
'inout' semantics) to the original parameter.
This commit introduces a new IR pass, 'undoParameterCopy', specifically
for CUDA/OptiX targets to address this. The pass operates as follows:
1. Identifies temporary IR variables created for 'inout' parameters,
which are now decorated with 'TempCallArgVarDecoration'.
2. Maps these temporary variables back to their original parameter
storage (e.g., the OptiX payload pointer).
3. Replaces all uses of the temporary variable directly with the
original parameter pointer.
4. Removes the temporary variable declaration and its initializing store
(which copied from the original parameter to the temporary).
By transforming the IR to operate directly on the original parameter
storage before any potential call to IgnoreHit(), this fix ensures
that all modifications are preserved, correctly resolving issue #6326.
The pass is integrated into the compilation flow for relevant targets.
* Refactor(IR): Optimize GetOptiXRayPayloadPtr for better DCE/CSE
To allow for more effective dead code elimination (DCE) and
common subexpression elimination (CSE) of `getOptiXRayPayloadPtr`
instructions, this commit:
- Marks `kIROp_GetOptiXRayPayloadPtr` as side-effect-free within
`IRInst::mightHaveSideEffects` (in `slang-ir.cpp`).
- Flags `GetOptiXRayPayloadPtr` as `HOISTABLE` in its definition
within `slang-ir-inst-defs.h`.
This addresses scenarios where multiple, potentially redundant,
calls to `getOptiXRayPayloadPtr` might appear in the IR,
allowing optimizers to produce cleaner and potentially more
efficient code for OptiX targets. This change supports efforts
to refine IR handling for ray-tracing shader stages.
* Remove debugging code
* Refactor UndoParameterCopyVisitor for improved performance
- Optimized IR traversal by combining multiple passes into a single scan
- Removed unnecessary dictionary, immediately replace uses when a temp var is found
- Reduced duplicate code paths by checking for both temp vars and redundant stores in one loop
- Better handling of the 'changed' flag to ensure DCE only runs when needed
- Results in fewer instruction traversals and improved efficiency for large functions
* Add Test
|
| |
|
| |
Fixes #7127
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add cluster geometry intrinsics for ray tracing
- Added GetClusterID() method to HitObject class
- Added CandidateClusterID() and CommittedClusterID() methods to
RayQuery class
- Added SPV_NV_cluster_acceleration_structure extension support
- Added GL_NV_cluster_acceleration_structure extension support
- Added test files for RayQuery and HitObject cluster methods
Fixes #6431
* OpRayQueryGetIntersectionClusterIdNV - unrecognized spirv
Disabling spirv backend for SPV_NV_cluster_acceleration_structure
hlsl.meta.slang(18674): error 29100: unrecognized spirv opcode:
OpRayQueryGetIntersectionClusterIdNV
result:$$int = OpRayQueryGetIntersectionClusterIdNV
&this $iCandidateOrCommitted;
^~~~~~
hlsl.meta.slang(18670): error 30019: expected an expression of type
'int', got 'void'
return spirv_asm
^~~~~~~~~
ninja: build stopped: subcommand failed.
* 6431 - Fix spirv opcode
* Remove tests
* Add relevant tests
* Review - Simplify tests
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* 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>
|
| |
|
|
|
| |
* Fix optix varying legalization.
* Add test.
|
| | |
|
| |
|
|
|
| |
* Fix CUDA prelude for makeMatrix
* Add regression test.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Handle out/inout functions with separate consideration
* Fixing bug with passing aliasable instructions
* Handle autodiff functions (fwd and rev) in warning system
* Handling interface methods
* Handling ref parameters like out/inout
* Temporary fix to remaining bugs
* Refactoring methods and tests
* Recursive check for empty structs
* Using default initializable interface in tests
* Resolving CI fail
|
| |
|
|
|
|
|
| |
* Remove use of `G0` and `__target_intrinsic` in stdlib.
* Fix.
* Fix calling intrinsic in global scope.
|
| |
|
|
|
|
|
|
|
| |
* Add `requirePrelude()` intrinsic function.
* Fix.
---------
Co-authored-by: Yong He <yhe@nvidia.com>
|
| |
|
| |
See https://github.com/shader-slang/slang/issues/2213
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* #include an absolute path didn't work - because paths were taken to always be relative.
* Add support for sizeOf/alignOf/offsetOf to stdlib.
Add $G intrinsic expansion that works of the generic parameters not the param type
* Test cuda layout.
* Fix CUDA layout issues.
Fix reflection to handle other built in types.
Fix __offsetOf
* Tests of reflection and layout as reported directly from CUDA.
* Comment about use of aligned size as size.
* Fix warning from VS.
* Check alignment is pow2.
* Small improvements to alignment calcs.
* Tab to spaces.
* Fix alignment pointer sizes on 32 bit OS for CUDA.
* Fix CUDA reflection on 32 bit.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* WIP: Trying to figure out how texturing will work with CUDA.
* WIP: Fixes for CUDA layout. Initial CUDA texture test.
* WIP: Outputs something compilable by CUDA for TextureND.Sample
* 2d texture working with CUDA.
* Fix how binding for SamplerState occurs in CUDA.
* Small tidy up of comments.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add test result for compile-to-cuda
* Add RAII for some CUDA types to simplify usage.
* First pass handling of some instrinsics on CUDA (for example transcendentals)
* CUDA working with built in intrinsics.
* Add missing CUDA prelude intrinsics.
* CUDA matches CPU output on simple-cross-compile.slang
* First pass at hlsl-scalar-float-intrinsic.slang test.
* Fix smoothstep impl on CUDA and CPU.
* Fixed step intrinsic on CUDA/CPU.
* Added operator[] to Matrix for C++, to allow row access.
Needs a fix for CUDA.
* Fixed warning on clang build.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* First pass at BindLocation.
* Added BindSet::init - for initializing with two input constant buffers. Needs better name, and perhaps should be another class.
* Fix handling of constant buffer stripping.
Improved initialization.
* Trying to generalize BindLocation a little more.
Split out CPULikeBindRoot.
* More work to make BindLocation et al work with non uniform bindings.
* Added parsing to a location.
* WIP: Trying to get CPU working with BindLocation.
* Describe problem of knowing the type of the reference point in the binding table.
* More ideas on getBindings fix.
* Remove BindSet as member of BindLocation.
* Added BindLocation::Invalid
* Made BindLocation able to be key in hash
* Use BindLocation for bindings on BindingSet.
* Added cuda and nvrtc categories to test infrastructure.
Disabled CUDA synthetic tests by default.
Fixed such that all tests now produce something in BindLocation style.
* Use m_userIndex instead of m_userData on Resource.
Move the binding setup out of cpu-compute-util (as no longer CPU specific)
* Removed CPUBinding - used BindLocation/BindSet instead.
Fixed some bugs around indexOf around uniform indirection.
* Renamed BindSet::Resource -> BindSet::Value.
* Document BindLocation.
* Fixes for Clang/GCC
Improve invariant requirement handling when constructing from BindPoints.
* WIP: First attempt to run CUDA kernel.
* Fix some issues around doing CUDA kernel launch.
* Fix issues around use of cudaMemCpy .
* Better cuda runtime error checking mechanism.
* Fixed bug in passing parameters to cuda kernel launch.
Simplified initialisation of context.
* WIP: Fix CUDA runtime issues.
* Add explicit CUDA synchronize so failures don't appear on implicit ones.
* Fix problem emitting non shared variable on CUDA.
* Fix some typos in CUDA layout.
Use just a pointer for now for CUDA StucturedBuffer.
* Arg order for CUDA launch was wrong.
* First compute kernel runs on CUDA.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* CPPCompiler -> DownstreamCompiler
* Added DownstreamCompileResult to start abstraction such that we don't need files.
* * Split out slang-blob.cpp
* Made CompileResult hold a DownstreamCompileResult - for access to binary or ISlangSharedLibrary
* Keep temporary files in scope.
* Add a hash to the hex dump stream.
* Move all file tracking into DownstreamCompiler.
* WIP support for nvrtc.
* WIP: Adding support for nvrtc compiler.
Adding enum types, wiring up the nvrtc into slang.
* Fix remaining CPPCompiler references.
* Fix order issue on target string matching.
* Use ISlangSharedLibrary for nvrtc.
* Use DownstreamCompiler for nvrtc.
* WIP first pass at compilation win nvrtc.
* Added testing if file is on file system into CommandLineDownstreamCompiler.
Added sourceContentsPath.
* Make test cuda-compile.cu work by just compiling not comparing output.
* Genearlize DownstreamCompiler usage.
* Fix warning on clang.
* Remove CompilerType from DownstreamCompiler.
* Use DownstreamCompiler interface for all compilers.
NOTE for FXC, DXC and GLSLANG this doesn't mean using 'compile' - it's still extracting functions from shared library.
* Replace DownstreamCompiler::SourceType -> SlangSourceLanguage
* Replace _canCompile with something data driven.
* Fix compiling on gcc/clang for DownstreamCompiler.
* Moved some text conversions into DownstreamCompiler.
* Fix problem on non-vc builds with not having return on locateCompilers for VS.
* Change so no warning for code not reachable on locateCompilers for vs.
* WIP: CUDA code generation - currently just using CPU layout and HLSL.
* emitXXXForEntryPoint -> emitEntryPointSource
emitSourceForEntryPoint -> emitEntryPointSourceFromIR
Fix up generating cuda to get PTX.
* WIP emitting cuda for IR.
* Small improvements to CUDA ouput.
* Disable the CUDA emit test, as output not currently compilable.
|
|
|
* CPPCompiler -> DownstreamCompiler
* Added DownstreamCompileResult to start abstraction such that we don't need files.
* * Split out slang-blob.cpp
* Made CompileResult hold a DownstreamCompileResult - for access to binary or ISlangSharedLibrary
* Keep temporary files in scope.
* Add a hash to the hex dump stream.
* Move all file tracking into DownstreamCompiler.
* WIP support for nvrtc.
* WIP: Adding support for nvrtc compiler.
Adding enum types, wiring up the nvrtc into slang.
* Fix remaining CPPCompiler references.
* Fix order issue on target string matching.
* Use ISlangSharedLibrary for nvrtc.
* Use DownstreamCompiler for nvrtc.
* WIP first pass at compilation win nvrtc.
* Added testing if file is on file system into CommandLineDownstreamCompiler.
Added sourceContentsPath.
* Make test cuda-compile.cu work by just compiling not comparing output.
* Fix warning on clang.
|
