diff options
| author | jsmall-nvidia <jsmall@nvidia.com> | 2020-01-29 10:06:45 -0500 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2020-01-29 10:06:45 -0500 |
| commit | 2c8b983cf20ba662e351813f3f432b65eef3530c (patch) | |
| tree | 61ee43c2d1db218f8f9deb91a83a7cee1a4340e7 /source/slang/slang-emit-cuda.cpp | |
| parent | 58cea79d8622a08b0887dbfda8f8042e42679c8f (diff) | |
Feature/fix cuda function preamble (#1187)
* Fix tests/compute/global-init.slang by handling some other cases where functions are emitted.
* Fix comment.
Diffstat (limited to 'source/slang/slang-emit-cuda.cpp')
| -rw-r--r-- | source/slang/slang-emit-cuda.cpp | 4 |
1 files changed, 1 insertions, 3 deletions
diff --git a/source/slang/slang-emit-cuda.cpp b/source/slang/slang-emit-cuda.cpp index 83ad4a0f8..2d49e606c 100644 --- a/source/slang/slang-emit-cuda.cpp +++ b/source/slang/slang-emit-cuda.cpp @@ -501,9 +501,7 @@ void CUDASourceEmitter::emitSimpleFuncParamsImpl(IRFunc* func) void CUDASourceEmitter::emitSimpleFuncImpl(IRFunc* func) { - // Mark as run on device. Don't need to worry about entry point, as that is output separately to call the __device_ implementation - m_writer->emit("__device__ "); - + // Skip the CPP impl - as it does some processing we don't need here for entry points. CLikeSourceEmitter::emitSimpleFuncImpl(func); } |
