Parallel Sparse Matrix Solver on the GPU Applied to Simulation of Electrical Machines

摘要

Nowadays, several industrial applications are being ported to parallelarchitectures. In fact, these platforms allow acquire more performance forsystem modelling and simulation. In the electric machines area, there are manyproblems which need speed-up on their solution. This paper examines theparallelism of sparse matrix solver on the graphics processors. Morespecifically, we implement the conjugate gradient technique with input matrixstored in CSR, and Symmetric CSR and CSC formats. This method is one of themost efficient iterative methods available for solving the finite-element basisfunctions of Maxwell's equations. The GPU (Graphics Processing Unit), which isused for its implementation, provides mechanisms to parallel the algorithm.Thus, it increases significantly the computation speed in relation to serialcode on CPU based systems.