Toward an efficient parallel eigensolver for dense symmetric matrices
- Sandia National Labs., Albuquerque, NM (United States). Applied and Numerical Math Dept.
- Univ. of Colorado, Boulder, CO (United States). Dept. of Computer Science
- Advanced Product and Process Technology, Tower Automotive, Milwaukee, WI (United States)
The authors describe a parallel algorithm for finding the eigenvalues and eigenvectors of a dense symmetric matrix, with an emphasis on the dense linear algebra operations. They follow the traditional three-step process: reduce to tridiagonal form, solve the tridiagonal problem, then back transform the result. Since the different steps have different algorithmic characteristics, this problem serves as a perfect vehicle for exploring some issues associated with parallel linear algebra calculations. In particular, they examine the effects of matrix distribution and blocking on the computational performance of tridiagonalization and backtransformation. Through experiments on an Intel Paragon, they demonstrate that block storage of the matrix is not necessary for a highly efficient block algorithm. The performance of the approach compares very favorably with that of the corresponding ScaLAPACK routines.
- Research Organization:
- Sandia National Laboratories (SNL), Albuquerque, NM, and Livermore, CA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC04-94AL85000; FG03-97ER25325
- OSTI ID:
- 355657
- Journal Information:
- SIAM Journal on Scientific Computing, Vol. 20, Issue 3; Other Information: PBD: Jan 1999
- Country of Publication:
- United States
- Language:
- English
Similar Records
The design of a parallel, dense linear algebra software library: Reduction to Hessenberg, tridiagonal, and bidiagonal form
Parallel performance of a symmetric eigensolver based on the Invariant Subspace Decomposition approach