Non-contiguous Processor Allocation Algorithms for Mesh-connected Multicomputers 1

Current processor allocation techniques for highly parallel systems are typically restricted to contiguous allocation strategies for which performance suuers signiicantly due to the inherent problem of fragmentation. As a result, message passing systems have yet to achieve the high utilization levels exhibited by traditional vector supercomputers. We are investigating processor allocation algorithms which lift the restriction on contiguity of processors in order to address the problem of fragmentation. Three non-contiguous processor allocation strategies: Paging allocation, Random allocation and the Multiple Buddy Strategy (MBS) are proposed and studied in this paper. Simulations compare the performance of the non-contiguous strategies with that of several well-known contiguous algorithms. We show that non-contiguous allocation algorithms perform better overall than the contiguous ones, even when message-passing contention is considered. We also present the results of experiments on an Intel Paragon XP/S-15 with 208 nodes that show non-contiguous allocation is feasible with current technologies.

[1]  E. Barszcz,et al.  Profiling the communication workload of an iPSC/860 , 1994, Proceedings of IEEE Scalable High Performance Computing Conference.

[2]  Yahui Zhu,et al.  Efficient Processor Allocation Strategie for Mesh-Connected Parallel Computers , 1992, J. Parallel Distributed Comput..

[3]  James L. Peterson,et al.  Buddy systems , 1977, CACM.

[4]  Melanie L. Fulgham.  Performance of Chaos and Oblivious Routers Under Non-uniform Traffic , 1993 .

[5]  J. Kim,et al.  A Top-Down Processor Allocation Scheme for Hypercube Computers , 1991, IEEE Trans. Parallel Distributed Syst..

[6]  Virginia Mary Lo,et al.  ProcSimity: an experimental tool for processor allocation and scheduling in highly parallel systems , 1995, Proceedings Frontiers '95. The Fifth Symposium on the Frontiers of Massively Parallel Computation.

[7]  Wei-Tek Tsai,et al.  Lookahead processor allocation in mesh-connected massively parallel multicomputer , 1994, Proceedings of 8th International Parallel Processing Symposium.

[8]  B. Bose,et al.  A new strategy for processors allocation in an N-cube multiprocessor , 1989, Eighth Annual International Phoenix Conference on Computers and Communications. 1989 Conference Proceedings.

[9]  V. Lo,et al.  Contiguous and Non-contiguous Processor Allocation , 1995 .

[10]  Ming-Syan Chen,et al.  Processor Allocation in an N-Cube Multiprocessor Using Gray Codes , 1987, IEEE Transactions on Computers.

[11]  John P. Hayes,et al.  Subcube Allocation in Hypercube Computers , 1991, IEEE Trans. Computers.

[12]  Mark S. Squillante,et al.  Performance analysis of job scheduling policies in parallel supercomputing environments , 1993, Supercomputing '93. Proceedings.

[13]  Keqin Li,et al.  A Two-Dimensional Buddy System for Dynamic Resource Allocation in a Partitionable Mesh Connected System , 1991, J. Parallel Distributed Comput..

[14]  Lionel M. Ni,et al.  A survey of wormhole routing techniques in direct networks , 1993, Computer.

[15]  Nian-Feng Tzeng,et al.  An efficient submesh allocation strategy for mesh computer systems , 1991, [1991] Proceedings. 11th International Conference on Distributed Computing Systems.

[16]  M. H. MacDougall.  Simulating computer systems: techniques and tools , 1989 .

[17]  Charles L. Seitz,et al.  The cosmic cube , 1985, CACM.

[18]  Chita R. Das,et al.  A Lazy Scheduling Scheme for Improving Hypercube Performance , 1993, 1993 International Conference on Parallel Processing - ICPP'93.

[19]  Bhagirath Narahari,et al.  Scheduling independent tasks on partitionable hypercube multiprocessors , 1993, [1993] Proceedings Seventh International Parallel Processing Symposium.

[20]  Bill Nitzberg,et al.  Non-contiguous processor allocation algorithms for distributed memory multicomputers , 1994, Proceedings of Supercomputing '94.

[21]  Phillip Krueger,et al.  ob Scheduling is More Important than Processor Allocation for Hypercube Computers , 1994, IEEE Trans. Parallel Distributed Syst..