Towards graph based parallel sparse solver for circuit simulation problems

In this paper, a novel technique based on graph theory is introduced to solve sparse linear systems. The proposed technique enhances the ability to build sparse parallel solvers for circuit simulation matrices. Thus the proposed technique improves the performance of circuit simulation algorithms. The new technique represents sparse linear system as a signal flow graph. Then it divides the graph into separate strongly connected components SCC. SCCs relations are represented and used to enhance the parallelism of the solver.