The distributed shared memory system is characterized by shared memory multi-processors on each node and distributed memory among nodes.In order to make use of this hierarchical architecture

this paper discusses the performance of MPI+OpenMP hybrid programming paradigm and different implementations.We design a multi-granularity parallel algorithm for solving larger scale tridiagonal linear systems

and compare its performance with pure MPI algorithm on the high performance computer of Shanghai University.The results indicate that the hybrid algorithm shows better speedup and scalability.