summaryrefslogtreecommitdiffstats
path: root/Help/release/dev/cuda-clang-separable-compilation.rst
Commit message (Collapse)AuthorAgeFilesLines
* CUDA: Clang separable compilationRaul Tambre2020-09-241-0/+4
For NVCC the compiler takes care of device linking when passed the "-dlink" flag. Clang doesn't support such magic and requires the buildsystem to do the work that NVCC does behind the scenes. The implementation is based on Bazel's device linking documentation: https://github.com/tensorflow/tensorflow/blob/7cabcdf073abad8c46e9dda62bb8fa4682d2061e/third_party/nccl/build_defs.bzl.tpl#L259 Closes: #20726