A distributed fair polling scheme applied to OR-parallel logic programming
- Univ. of Maryland, College Park (United States)
A task scheduling algorithm for parallel execution of logic programs on NUMA multiprocessors is proposed. The algorithm endorces a so-called fair polling policy. The authors show analytically that the proposed algorithm has a good iso-efficiency function. Results from simulation on switch based NUMA architecture multiprocessors, the BBN Butterfly GP1000 and TC2000, corroborate the analysis. The proposed algorithm exhibits performance characteristics very similar to that of its counterpart that uses a shared memory. It achieves reasonable speed-up on benchmarks, using a nonconstant time task migration protocol. In addition, fair polling algorithms (with or without a shared memory) are shown to be consistently superior than several other known polling schemes that do not maintain fairness, for a variety of benchmark programs.
- OSTI ID:
- 6987949
- Journal Information:
- International Journal of Parallel Programming; (United States), Journal Name: International Journal of Parallel Programming; (United States) Vol. 20:6; ISSN IJPPE5; ISSN 0885-7458
- Country of Publication:
- United States
- Language:
- English
Similar Records
Performance of an OR-parallel logic programming system
A task adaptive parallel graphics rendering algorithm