An Evaluation of Medium - Grain Data

In this paper, we study several issues related to the medium grain dataaow model of execution. We present bottom-up compilation of medium grain clusters from a ne grain dataaow graph. We compare the basic block and the dependence sets algorithms that partition dataaow graphs into clusters. For an extensive set of benchmarks we assess the average number of instructions in a cluster and the reduction in matching operations compared with ne grain dataaow execution. We study the performance of medium grain dataaow when several architectural parameters, such as the number of processors, matching cost, and network latency, are varied. The results indicate that medium grain execution ooers a good speedup over the ne grain model, that it is scalable, and tolerates network latency and high matching costs well. Medium grain execution can beneet from a higher output bandwidth of a processor and nally, a simple superscalar processor with an issue rate of two is suucient to exploit the internal parallelism of a cluster.