projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: b0b1f45) | patch
[CUDA][NFC] Split math.h functions out of __clang_cuda_device_functions.h
authorJohannes Doerfert <johannes@jdoerfert.de>
Thu, 2 Apr 2020 03:22:13 +0000 (22:22 -0500)
committerJohannes Doerfert <johannes@jdoerfert.de>
Thu, 2 Apr 2020 04:46:27 +0000 (23:46 -0500)
commitd1705c1196fedfe927716923ac121f1134924a36
tree7b78e566b952f34f4528129e960c59dcf0220142tree | snapshot
parentb0b1f451ae693703acf7fb73ba5a772aaa070d67commit | diff
[CUDA][NFC] Split math.h functions out of __clang_cuda_device_functions.h

This is not supported to change anything but allow us to reuse the math
functions separately from the device functions, e.g., source them at
different times. This will be used by the OpenMP overlay.

This also adds two `return` keywords that were missing.

Reviewed By: tra

Differential Revision: https://reviews.llvm.org/D77238
clang/lib/Headers/CMakeLists.txt diff | blob | history
clang/lib/Headers/__clang_cuda_device_functions.h diff | blob | history
clang/lib/Headers/__clang_cuda_math.h [new file with mode: 0644] blob
clang/lib/Headers/__clang_cuda_runtime_wrapper.h diff | blob | history
Domain: System / Toolchain;