Optimal eigenvalue computation on distributed-memory MIMD multiprocessors
Simon proves that bisection is not the optimal method for computing an eigenvalue on a single vector processor. In this paper, we show that his analysis does not extend in a straightforward way to the computation of an eigenvalue on a distributed-memory MIMD multiprocessor. In particular, we show how the optimal number of sections (and processors) to use for multisection depends on variables such as the matrix size and certain parameters inherent to the machine. We also show that parallel multisection outperforms the variant of parallel bisection proposed by Swarztrauber or this problem on a distributed-memory MIMD multiprocessor. We present the results of experiments on the 64-processor Intel iPSC/2 hypercube and the 512-processor Intel Touchstone Delta mesh multiprocessor.
- Research Organization:
- Colorado Univ., Boulder, CO (United States). Dept. of Computer Science
- Sponsoring Organization:
- USDOE, Washington, DC (United States); National Science Foundation, Washington, DC (United States)
- DOE Contract Number:
- FG02-92ER25122; AC05-84OR21400
- OSTI ID:
- 10177289
- Report Number(s):
- DOE/ER/25122-4; CU-CS-617-92; CONF-930331-14; ON: DE93018337; CNN: Grant CCR-9109785
- Resource Relation:
- Conference: 6. Society for Industrial and Applied Mathematics (SIAM) conference on parallel processing for scientific computing,Norfolk, VA (United States),21-24 Mar 1993,; Other Information: PBD: Oct 1992
- Country of Publication:
- United States
- Language:
- English
Similar Records
Solving the symmetric tridiagonal eigenvalue problem on the hypercube
Multiprocessor execution of functional programs