Using Hybrid Parallel Programming Techniques for the Computation, Assembly and Solution Stages in Finite Element Codes

The so called “hybrid parallelism paradigm”, that combines programming techniques for architectures with distributed and shared memories using MPI (Message Passing Interface) and OpenMP (Open Multi-Processing) standards, is currently adopted to exploit the growing use of multi-core computers, thus improving the efficiency of codes in such architectures (several multi-core nodes or clustered symmetric multi-processors (SMP) connected by a fast net to do exhaustive computations). In this paper a parallel hybrid finite element code is developed and its performance is evaluated, using MPI for communication between cluster nodes and OpenMP for parallelism within the SMP nodes. An efficient thread-safe matrix library for computing element/cell residuals (or right hand sides) and Jacobians (or matrices) in FEM-like codes is introduced and fully described. The cluster in which the code was tested is the CIMEC’s ‘Coyote’ cluster, which consists of eight-core computing nodes connected through Gigabit Ethernet. Keywords— Finite Elements, MPI, OpenMP, PETSc, hybrid programming, Matrix Library.