Analysis of the impact of memory in distributed parallel processing systems
SIGMETRICS '94 Proceedings of the 1994 ACM SIGMETRICS conference on Measurement and modeling of computer systems
Evaluation of design choices for gang scheduling using distributed hierarchical control
Journal of Parallel and Distributed Computing
An evaluation of parallel job scheduling for ASCI Blue-Pacific
SC '99 Proceedings of the 1999 ACM/IEEE conference on Supercomputing
Paging tradeoffs in distributed-shared-memory multiprocessors
Proceedings of the 1994 ACM/IEEE conference on Supercomputing
Packing Schemes for Gang Scheduling
IPPS '96 Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing
Memory Usage in the LANL CM-5 Workload
IPPS '97 Proceedings of the Job Scheduling Strategies for Parallel Processing
Improved Utilization and Responsiveness with Gang Scheduling
IPPS '97 Proceedings of the Job Scheduling Strategies for Parallel Processing
Improving First-Come-First-Serve Job Scheduling by Gang Scheduling
IPPS/SPDP '98 Proceedings of the Workshop on Job Scheduling Strategies for Parallel Processing
Gang Scheduling with Memory Considerations
IPDPS '00 Proceedings of the 14th International Symposium on Parallel and Distributed Processing
Utilization and Predictability in Scheduling the IBM SP2 with Backfilling
IPPS '98 Proceedings of the 12th. International Parallel Processing Symposium on International Parallel Processing Symposium
| Hi-index | 0.00 |
The addition of time-slicing to space-shared gang scheduling improves the average response time of the jobs in a typical job stream. Recent research has shown that time-slicing is most effective when the jobs admitted for execution fit entirely into physical memory. The question is, how to select and map jobs to make the best use of the available physical memory. Specifically, the achievable degree of multi-programming is limited by the memory requirements, or physical memory pressure, of the admitted jobs. We investigate two techniques for improving the performance of gang scheduling in the presence of memory pressure: 1) a novel backfill approach which improves memory utilization, and 2) an adaptive multi-programming level which balances processor/memory utilization with job response time performance. Our simulations show that these techniques reduce the average wait time and slow-down performance metrics over naive first-come-first-serve methods on a distributed memory parallel system.