Auto-generation of Parallel Finite-Differencing Code for MPI, TBB and CUDA | IEEE Conference Publication | IEEE Xplore