diff options
| author | jsmall-nvidia <jsmall@nvidia.com> | 2021-06-10 14:57:09 -0400 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2021-06-10 11:57:09 -0700 |
| commit | 37e8917d10626b519470f2e34625f0efe741352f (patch) | |
| tree | 4e8e51bd63ebc03fcdf0c893b675906cf507d73c /tests/cuda/cuda-reflection.slang | |
| parent | 0d9bd79e8fd4d57e1a723ca6b6a45efec2b42872 (diff) | |
CUDA layout corner cases/testing (#1881)
* #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.
Diffstat (limited to 'tests/cuda/cuda-reflection.slang')
| -rw-r--r-- | tests/cuda/cuda-reflection.slang | 28 |
1 files changed, 28 insertions, 0 deletions
diff --git a/tests/cuda/cuda-reflection.slang b/tests/cuda/cuda-reflection.slang new file mode 100644 index 000000000..95bf591c9 --- /dev/null +++ b/tests/cuda/cuda-reflection.slang @@ -0,0 +1,28 @@ +// cuda-reflection.slang + +//TEST:REFLECTION:-stage compute -entry main -target cuda + +struct PadLadenStruct +{ + double a; + uint8_t b; +}; + +// This is to check if the last half can be inserted 'inside' the spare padding of a. It should not be +struct StructWithArray +{ + PadLadenStruct a[1]; + uint8_t c; + + matrix<half, 3, 3> d; + uint8_t e; +}; + +ConstantBuffer<StructWithArray> cb; +RWStructuredBuffer<StructWithArray> sb; + +[numthreads(1, 1, 1)] +void main( + uint3 dispatchThreadID : SV_DispatchThreadID) +{ +}
\ No newline at end of file |
