HIGH-PERFORMANCE PCG SOLVERS FOR FEM STRUCTURAL-ANALYSIS

The preconditioned conjugate gradient algorithm is a well-known and po werful method used to solve large sparse symmetric positive definite l inear systems. Such systems are generated by the finite element discre tization in structural analysis but users of finite elements in this c ontext generally still rely on direct methods. It is our purpose in th e present work to highlight the improvement brought forward by some ne w preconditioning techniques and show that the preconditioned conjugat e gradient method performs better than efficient direct methods.