A parallel Jacobi-Davidson algorithm with Block FSAI preconditioning

Preconditioning is a key factor to accelerate the convergence of sparse eigensolvers. The present communication describes the acceleration of a parallel Jacobi-Davidson algorithm by using Block FSAI (BFSAI) as a preconditioner for symmetric positive definite eigenproblems. BFSAI proves a robust and efficient preconditioner in a number of test cases arising from the Finite Element discretization of 3D fluid-dynamical and mechanical real engineering applications