Cleaned up CUDA compilation
commitdd669b1b6fb0b574ae4def338cbabc0e24152bc6
authorMark Abraham <mark.j.abraham@gmail.com>
Fri, 2 Oct 2020 15:44:59 +0000 (2 15:44 +0000)
committerPaul Bauer <paul.bauer.q@gmail.com>
Fri, 2 Oct 2020 15:44:59 +0000 (2 15:44 +0000)
tree1f8fc91ea925a1ce63381c35070e213374de9a66
parentfff5715e08cf6cfcada830806c792bf5495a0253
Cleaned up CUDA compilation

Some .cpp source files used CUDA run-time headers that need
special handling when e.g. clang issues warnings from antiquated
constructs in older NVIDIA runtime headers. This is now handled
centrally in a similar manner to that required for SYCL compilation.

Dependency of the higher-level taskassignment module on GPU
API-specific compilation was removed, instead deploying some new
helper functions.
src/gromacs/CMakeLists.txt
src/gromacs/ewald/CMakeLists.txt
src/gromacs/gpu_utils/CMakeLists.txt
src/gromacs/gpu_utils/devicebuffer.cuh
src/gromacs/hardware/CMakeLists.txt
src/gromacs/hardware/device_management.h
src/gromacs/hardware/device_management_common.cpp
src/gromacs/mdtypes/CMakeLists.txt
src/gromacs/nbnxm/CMakeLists.txt
src/gromacs/taskassignment/CMakeLists.txt
src/gromacs/taskassignment/usergpuids.cpp