MPI Process and Network Device Affinitization for Optimal HPC Application Performance | IEEE Conference Publication | IEEE Xplore