This report describes and analyzes numerical experiments carried out with various symmetric eigensolvers in the context of the material science code Wien 97. Of particular interest are the performance improvements achieved with a new Level 3 eigensolver. The techniques which lead to a significant speed up are (1) sophisticated blocking in the tridiagonalization step, which leads to a twosweep algorithm, and (2) computing eigenvectors using inverse iteration on an appropriately chosen band matrix. The new Level 3 eigensolver improves the locality of data references and leads to performance improvements of up to 200 % (depending on hardware features and the problem size)
AbstractSolving dense symmetric eigenvalue problems and computing singular value decompositions cont...
An efficient parallel algorithm, farmzeroinNR, for the eigenvalue problem of a symmetric tridiagonal...
We present new algorithms that accelerate the bisection method for the symmetric tridiagonal eigenva...
We investigate the performance of the routines in LAPACK and the Successive Band Reduction (SBR) too...
The solution of the symmetric eigenvalue problem is a compute-intensive task in many scientific and ...
AbstractWe compare different algorithms for computing eigenvalues and eigenvectors of a symmetric ba...
In this report a way to apply high level Blas to the tridiagonalization process of a symmetric matri...
Complex symmetric matrices often appear in quantum physics in the solution methods of partial differ...
In the first-principles calculation of electronic structures, one of the most timeconsuming tasks is...
LAPACK is often mentioned as a positive example of a software library that encapsulates complex, rob...
. In this paper, we present preliminary results on a complete eigensolver based on the Yau and Lu me...
The solution of (generalized) eigenvalue problems for symmetric or Hermitian matrices is a common su...
We investigate the performance of the routines in LAPACK and the Successive Band Reduction (SBR) too...
This report demonstrates parallel versions of the Eispack functions TRED2 and TQL2 for finding all...
In the nonsymmetric eigenvalue problem, work has focused on the Hessenberg reduction and QR iteratio...
AbstractSolving dense symmetric eigenvalue problems and computing singular value decompositions cont...
An efficient parallel algorithm, farmzeroinNR, for the eigenvalue problem of a symmetric tridiagonal...
We present new algorithms that accelerate the bisection method for the symmetric tridiagonal eigenva...
We investigate the performance of the routines in LAPACK and the Successive Band Reduction (SBR) too...
The solution of the symmetric eigenvalue problem is a compute-intensive task in many scientific and ...
AbstractWe compare different algorithms for computing eigenvalues and eigenvectors of a symmetric ba...
In this report a way to apply high level Blas to the tridiagonalization process of a symmetric matri...
Complex symmetric matrices often appear in quantum physics in the solution methods of partial differ...
In the first-principles calculation of electronic structures, one of the most timeconsuming tasks is...
LAPACK is often mentioned as a positive example of a software library that encapsulates complex, rob...
. In this paper, we present preliminary results on a complete eigensolver based on the Yau and Lu me...
The solution of (generalized) eigenvalue problems for symmetric or Hermitian matrices is a common su...
We investigate the performance of the routines in LAPACK and the Successive Band Reduction (SBR) too...
This report demonstrates parallel versions of the Eispack functions TRED2 and TQL2 for finding all...
In the nonsymmetric eigenvalue problem, work has focused on the Hessenberg reduction and QR iteratio...
AbstractSolving dense symmetric eigenvalue problems and computing singular value decompositions cont...
An efficient parallel algorithm, farmzeroinNR, for the eigenvalue problem of a symmetric tridiagonal...
We present new algorithms that accelerate the bisection method for the symmetric tridiagonal eigenva...