An Assessment of Assignment Schemes for Dependency Graphs

Abstract The recent past has seen a number of publications proposing assignment schemes that do not ignore communication costs. Yet, little effort has been made to comparatively assess the performance and end-results of these schemes. This paper presents a comparative assessment of assignment schemes for dependency graphs that take into account the communication costs. To generate an assignment, some parameters describing the program model will be required. In many cases, accurate estimation of these parameters is hard. It is thought that inaccuracies in the estimation would lead to poor assignments. This paper investigates this speculation and presents experimental evidence that shows such inaccuracies do not greatly affect the quality of the assignments.

[1]  Nicola Cabibbo,et al.  A new method for updating SU(N) matrices in computer simulations of gauge theories , 1982 .

[2]  George Cybenko,et al.  Supercomputer performance evaluation and the Perfect Benchmarks , 1990, ICS '90.

[3]  Peter Thanisch,et al.  Assigning dependency graphs onto processor networks , 1991, Parallel Comput..

[4]  Edward G. Coffman,et al.  Computer and job-shop scheduling theory , 1976 .

[5]  Behrooz Shirazi,et al.  Analysis and Evaluation of Heuristic Methods for Static Task Scheduling , 1990, J. Parallel Distributed Comput..

[6]  Victor J. Rayward-Smith,et al.  UET scheduling with unit interprocessor communication delays , 1987, Discret. Appl. Math..

[7]  S. Manoharan.  Genesis: a generic simulation subsystem for parallel architectures , 1992, CompEuro 1992 Proceedings Computer Systems and Software Engineering.

[8]  Frank D. Anger,et al.  Scheduling Precedence Graphs in Systems with Interprocessor Communication Times , 1989, SIAM J. Comput..

[9]  Geoffrey C. Fox,et al.  The Perfect Club Benchmarks: Effective Performance Evaluation of Supercomputers , 1989, Int. J. High Perform. Comput. Appl..

[10]  K. Mani Chandy,et al.  A comparison of list schedules for parallel processing systems , 1974, Commun. ACM.

[11]  Peter Winkler,et al.  Counting linear extensions is #P-complete , 1991, STOC '91.

[12]  Hesham El-Rewini,et al.  Scheduling Parallel Program Tasks onto Arbitrary Target Machines , 1990, J. Parallel Distributed Comput..

[13]  Leslie G. Valiant,et al.  The Complexity of Computing the Permanent , 1979, Theor. Comput. Sci..

[14]  Jing-Jang Hwang,et al.  Multiprocessor scheduling with interprocessor communication delays , 1988 .