RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación
Contribución de congreso (ConferenceOutput)
A Parallel Variant of the Gram-Schmidt Process with Reorthogonalization
[EN] This work is concerned with the effective parallelization of algorithms for computing an orthonormal basis of a subspace in the context of Krylov-type eigenvalue solvers such as Arnoldi. The main
motivation is to try to improve the parallel efficiency of eigensolvers implemented in SLEPc, the
Scalable Library for Eigenvalue Problem Computations [4], especially when addressing large-scale
eigenproblems with many processors in distributed memory platforms., Part of this research used resources of the National Energy Research Scientific
Computing Center, which is supported by the Office of Science of the U.S. Department of Energy
under Contract No. DE-AC03-76SF00098.