Performance of a Benchmark Implementation of the Van Slyke and Wets Algorithm for Stochastic Programs on the Alliant FX/8
- Argonne National Laboratory (ANL), Argonne, IL (United States); Washington State Univ., Pullman, WA (United States)
A benchmark parallel implementation of the Van Slyke and Wets algorithm for stochastic linear programs, and the results of a carefully designed numerical experiment on the Sequent/Balance using the implementation are presented. An important use of this implementation is as a benchmark to assess the performance of approximation algorithms for stochastic linear programs. These approximation algorithms are best suited for implementation on parallel vector processes like the Alliant FX/8. Therefore, the performance of the benchmark implementation on the Alliant FX/8 is of interest. In this paper, we present results observed when a portion of the numerical experiment is performed on the Alliant FX/8. These results indicate that the implementation makes satisfactory use of the concurrency capabilities of the Alliant FX/8. They also indicate that the vectorization capabilities of the Alliant FX/8 are not satisfactorily utilized by the implementation.
- Research Organization:
- Argonne National Laboratory (ANL), Argonne, IL (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC); National Science Foundation (NSF)
- DOE Contract Number:
- FG06-87ER25045; W-31109-ENG-38
- OSTI ID:
- 5498553
- Report Number(s):
- ANL/CP--73599; CONF-910326--6; ON: DE91016515; CNN: DMS-8918785
- Country of Publication:
- United States
- Language:
- English
Similar Records
Performance of the Lawrence Livermore National Lab benchmark on the Alliant FX/8
Vector processing on the Alliant FX/8 multiprocessor