Download Now Free registration required
Optimal network performance is critical to efficient parallel scaling for communication-bound applications on large machines. With wormhole routing, no-load latencies do not increase significantly with number of hops traveled. Yet, the authors and others have recently shown that in presence of contention, message latencies can grow substantially large. Hence task mapping strategies should take the topology of the machine into account on large machines. In this paper, they present topology aware mapping as a technique to optimize communication on 3-dimensional mesh interconnects and hence improve performance. Their methodology builds upon the idea of object-based decomposition used in Charm++ which separates the processes of decomposition from mapping of computation to processors and allows a more flexible mapping based on communication patterns between objects.
- Format: PDF
- Size: 356.76 KB