ABSTRACT
When multiple jobs compete for processing resources on a parallel computer, the operating system kernel's processor allocation policy determines how many and which processors to allocate to each. In this paper we investigate the issues involved in constructing a processor allocation policy for large scale, message-passing parallel computers supporting a scientific workload.
We make four specific contributions:
We define the concept of efficiency preservation as a characteristic of processor allocation policies. Efficiency preservation is the degree to which the decisions of the processor allocator degrade the processor efficiencies experienced by individual applications relative to their efficiencies when run alone.
We identify the interplay between the kernel processor allocation policy and the application load distribution policy as a determinant of efficiency preservation.
We specify the details of two families of processor allocation policies, called Equipartition and Folding. Within each family, different member policies cover a range of efficiency preservation values, from very high to very low.
By comparing policies within each family as well as between families, we show that high efficiency preservation is essential to good performance, and that efficiency preservation is a more dominant factor in obtaining good performance than is equality of resource allocation.
- 1.I. Ashok. Adhara: A Run-Time Support System for @ace-Based Applications. PhD thesis, University of Washington, In Preparation.Google Scholar
- 2.M.-S. Chen and K.G. Shin. Processor allocation in an N-cube multiprocessor using gray codes. IEEE Transactions on Computers, C-36(12):1396-1407, December 1987. Google ScholarDigital Library
- 3.R. Cypher, A. I-Io, S. Konstantinidou, and P. Messina. Architectural requirements of parallel scientifc applications with explicit communication. In Proceedings 20th Annual International Symposium on Computer Architecture, pages 2-13, May 1993. Google ScholarDigital Library
- 4.K. Dussa, B. Carlson, L. Dowdy, and K-H. Park. Dynamic partitioning in a transputer environment. In Proceedings of A CM SIGMETRICS Conference, pages 203-213, May 1990. Google ScholarDigital Library
- 5.D.G. Feitelson. In Support of Gang Scheduling. PhD thesis, Department of Computer Science, The Hebrew University, December 1991.Google Scholar
- 6.D.G. Feitelson and L. Rudolph. Distributed hierarchical control for parallel processing. Computer, 23(5):65-77, May 1990. Google ScholarDigital Library
- 7.A. Gupta, A. Tucker, and S. Urushibara. The impact of operating system scheduling policies and synchronization methods on the performance of parallel applications. In Proceedings of A CM SIGMETRICS Conference, pages 120-132, May 1991. Google ScholarDigital Library
- 8.S. Leutenegger and M. Vernon. The performance of multiprogrammed multiprocessor scheduling policies. in Procee&ngs of A CM SIGMETRICS Conference, pages 226-236, May 1990. Google Scholar
- 9.S. Majumdar, D.L. Eager, and R. Bunt. Scheduling in multiprogrammed parallel systems. In Proceedings of ACM SIGMETRICS Conference, pages 104-113, May 1988. Google ScholarDigital Library
- 10.C. McCann, R. Vaswani, and J. Zahorjan. A dynamic processor allocation policy for multiprogrammed, shared memory multiprocessors. A CM Transactions on Computer Systems, 11(2):146-178, May 1993. Google ScholarDigital Library
- 11.D.M. Nicol and J.C. Townsend. Accurate modeling of parallel scientific computations. In Proceedings of A CM SIGMETRICS Conference, pages 165-170, May 1989. Google ScholarDigital Library
- 12.J. Ousterhout. Scheduling techniques for concurrent systems. In 3rd internatzonal Conference on D~stributed Computing Systems, pages 22-30, October 1982.Google Scholar
- 13.S. Setia, M.S. Squillante,, and S. Tripathi. Processor scheduling on multiprogrammed, distributed memory parallel systems. In Proceedings of A CM SIGMETRICS Conference, pages 158-170, May 1993. Google ScholarDigital Library
- 14.K.C. Sevcik. Characterization of parallelism in applications and their use in scheduling. In Procee&ngs of ACM SIGMETRICS Conference, pages 171-180, May 1989. Google Scholar
- 15.K.C. Sevcik. Application scheduling and processor allocation in multiprogrammed parallel processing systems. Performance Evaluation, To appear. Google ScholarDigital Library
- 16.M.S. Squillante and E.D. Lazowska. Using processorcache affinity information in shared-memory multiprocessor scheduling. IEEE Transactions on Parallel and D~str~buted Systems, 4(2):131-143, February 1993. Google ScholarDigital Library
- 17.C.A. Thekkath and H.M. Levy. Limits to low-latency communication on high-speed networks. ACM Transactions on Computer Systems, 11(2):179-203, May 1993. Google ScholarDigital Library
- 18.A. Tucker and A. Gupta. Process control and scheduling issues for multiprogrammed shared-memory multiprocessors. In Proceedings of the 12th A CM Symposzum on Operating System Pr~nc,ples, pages 159-166, December 1989. Google ScholarDigital Library
- 19.R. Vaswani and J. Zahorjan. The implications of cache affinity on processor scheduling for multiprogrammed, shared memory multiprocessors. In Proceedings 13th A CM Symposium on Operating Systems Prone,pies, pages 26-40, October 1991. Google ScholarDigital Library
- 20.T. von Eicken, D.E. Culler, S.C. Goldstein, and K.E. Schauser. Active messages: A mechanism for integrated communication and computation. In Proceed,ngs 19th International Symposium on Computer Arch,tecture, pages 256-266, May 1992 Google ScholarDigital Library
- 21.J. Zahorjan and C. McCann. Processor scheduling in shared memory multiprocessors. In Procee&ngs of A CM SIGMETRICS Conference, pages 214-225, M~y 1990. Google ScholarDigital Library
- 22.S. Zhou and T. Brecht. Processor-pool-based scheduling for large-scale NUMA multiprocessors. In Proceedings of A CM SIGMETRICS Conference, pages 133- 142, May 1991. Google ScholarDigital Library
Index Terms
- Processor allocation policies for message-passing parallel computers
Recommendations
Processor allocation policies for message-passing parallel computers
When multiple jobs compete for processing resources on a parallel computer, the operating system kernel's processor allocation policy determines how many and which processors to allocate to each. In this paper we investigate the issues involved in ...
Performance-Driven Processor Allocation
In current multiprogrammed multiprocessor systems, to take into account the performance of parallel applications is critical to decide an efficient processor allocation. In this paper, we present the Performance-Driven Processor Allocation policy (PDPA)...
Comments