Some parallel version of push label algorithm
in mpi_impl floder
run:
mpiexec -n <number of processes> ./mpi_push_relabel <input file>
in pthread_impl floder
run:
./pthread_push_relabel <input file> <number of threads>
in cuda_impl floder
run:
./cuda_push_relabel <input file> <num of blocks per grid> <number of thread per block>