ocl: separated/modularized support for FP-atomics #1461

Job	Run time
pre-commit	1m 4s
build-on-opencl (OPENMP=ON, SMM=libxsmm)	7m 47s
build-on-cuda (OPENMP=ON)	7m 46s
build-and-test (MPI=ON, OPENMP=ON, SMM=blas, openmpi)	4m 27s
build-on-rocm (OPENMP=ON)	8m 15s
build-and-test (MPI=ON, OPENMP=ON, SMM=blas, mpich)	3m 48s
build-and-test (MPI=ON, OPENMP=ON, SMM=libxsmm, openmpi)	3m 57s
build-and-test (MPI=ON, OPENMP=ON, SMM=libxsmm, mpich)	3m 55s
build-and-test (MPI=ON, OPENMP=OFF, SMM=blas, openmpi)	9m 7s
build-and-test (MPI=ON, OPENMP=OFF, SMM=blas, mpich)	9m 21s
build-and-test (MPI=ON, OPENMP=OFF, SMM=libxsmm, openmpi)	9m 7s
build-and-test (MPI=ON, OPENMP=OFF, SMM=libxsmm, mpich)	9m 13s
build-and-test (MPI=OFF, OPENMP=ON, SMM=blas, openmpi)	5m 23s
build-and-test (MPI=OFF, OPENMP=ON, SMM=libxsmm, openmpi)	5m 13s
build-and-test (MPI=OFF, OPENMP=OFF, SMM=blas, openmpi)	3m 27s
build-and-test (MPI=OFF, OPENMP=OFF, SMM=libxsmm, openmpi)	3m 13s
Combine & check coverage.	45s
	1h 35m 48s

Provide feedback