Dynamic remapping of parallel computations with varying resource demands
A large class of computational problems are characterized by frequent synchronization, and computational requirements which change as a function of time. When such a problem is solved on a message passing multiprocessor machine, the combination of these characteristics leads to system performance which deteriorates in time. Performance can be improved with periodic redistribution of computational load; however, redistribution can exact a sometimes large delay cost. The authors study the issue of deciding when to invoke a global load remapping mechanism. Such a decision policy must effectively weigh the costs of remapping against the performance benefits, and should be general enough to apply automatically to a wide range of computations. The authors propose a general remapping decision heuristic, then study its effectiveness and its anticipated behavior on two very different models of load evolution. Assuming only that the remapping cost is know, this policy dynamically attempts to minimize system degradation including the cost of remapping) per computation step. This policy is quite simple, choosing to remap when the first local minimum in the degradation function is detected.
- Research Organization:
- Dept. of Computer Science, College of William and Mary, Williamsburg, VA (US)
- OSTI ID:
- 6201257
- Journal Information:
- IEEE Trans. Comput.; (United States), Vol. 37:9
- Country of Publication:
- United States
- Language:
- English
Similar Records
Optimal dynamic remappping of data parallel computations
Parallel tetrahedral mesh adaptation with dynamic load balancing