WIP: Add GPU aware MPI support in cannon algorithm #1307

Job	Run time
pre-commit	1m 31s
build-on-opencl (OPENMP=ON, SMM=libxsmm)	11m 43s
build-on-cuda (OPENMP=ON)	14m 31s
build-and-test (MPI=ON, OPENMP=ON, SMM=blas, openmpi)	2m 31s
build-and-test (MPI=ON, OPENMP=ON, SMM=blas, mpich)	2m 32s
build-and-test (MPI=ON, OPENMP=ON, SMM=libxsmm, openmpi)	2m 29s
build-and-test (MPI=ON, OPENMP=ON, SMM=libxsmm, mpich)	2m 24s
build-and-test (MPI=ON, OPENMP=OFF, SMM=blas, openmpi)	2m 29s
build-and-test (MPI=ON, OPENMP=OFF, SMM=blas, mpich)	2m 29s
build-and-test (MPI=ON, OPENMP=OFF, SMM=libxsmm, openmpi)	2m 32s
build-and-test (MPI=ON, OPENMP=OFF, SMM=libxsmm, mpich)	2m 31s
build-and-test (MPI=OFF, OPENMP=ON, SMM=blas, openmpi)	2m 31s
build-and-test (MPI=OFF, OPENMP=ON, SMM=libxsmm, openmpi)	2m 29s
build-and-test (MPI=OFF, OPENMP=OFF, SMM=blas, openmpi)	2m 27s
build-and-test (MPI=OFF, OPENMP=OFF, SMM=libxsmm, openmpi)	2m 30s
build-on-rocm (OPENMP=ON)	4m 59s
Combine & check coverage.	0s
	1h 2m 38s

Provide feedback