llvm
f7a9fcad - [flang][cuda] Use PTX instruction for atomicAdd with 4xf32 (#169581)

Commit

127 days ago

[flang][cuda] Use PTX instruction for atomicAdd with 4xf32 (#169581) Implementation similar to the clang one in `clang/lib/Headers/__clang_cuda_intrinsics.h`

References

#20851 - LLVM and SPIRV-LLVM-Translator pulldown (WW49 2025)

Author

clementval

Parents

fd22706e

llvm f7a9fcad - [flang][cuda] Use PTX instruction for atomicAdd with 4xf32 (#169581)

llvm
f7a9fcad - [flang][cuda] Use PTX instruction for atomicAdd with 4xf32 (#169581)