Vectorization of a multiprocessor multifrontal code
- CERFACS, Toulouse (FR)
- Harwell Lab., Oxfordshire (GB)
The authors describe design changes that enhance the vectorization of a multiprocessor version of a multifrontal code for the direct solution of large sparse sets of linear equations. These changes employ techniques used with success in full Gaussian elimination and are based on the use of matrix-vector and matrix-matrix kernels as implemented in the Level 2 and Level 3 BLAS. They illustrate the performance of the improved code by runs on the IBM 3090/VF, the ETA-10P, and the CRAY-2. Although their experiments are principally on a single processor of these machines, they briefly consider the influence of multiprocessing. Speedup factors of more than 11 are obtained, and the modified code performs at over 200 MFLOPS on standard structures problems on one processor of the CRAY-2.
- OSTI ID:
- 5422748
- Journal Information:
- International Journal of Supercomputer Application; (USA), Vol. 3:3; ISSN 0890-2720
- Country of Publication:
- United States
- Language:
- English
Similar Records
Use of Level 3 BLAS in LU factorization in a multiprocessing environment on three vector multiprocessors: The ALLIANT FX/80, the CRAY-2, and the IBM 3090 VF
Use of level 3 BLAS in lu factorization in a multiprocessing environment on three vector multiprocessors; The Alliant FX/80, the CRAY-2, and the IBM 3090 VF