Parallel rapid operator for iterative finite element solvers on a shared memory machine (Q1208497)

The authors consider the \(p\)-version of the finite element method and suggest the use of iterative methods for solving the resulting linear system of algebraic equations. They describe an efficient way of performing matrix-vector multiplication (which is at the heart of any iterative solver) which minimizes storage requirements by eliminating the assembly of global and element matrices and which is well suited for implementation on parallel and vector processor, shared memory computers. Finally they exhibit results from numerical experiments showing the efficiency of the proposed scheme.