Reducing Power with Performance Constraints for Parallel Sparse Applications

Sparse and irregular computations constitute a large fraction of applications in the data-intensive scientific do-main. While every effort is made to balance the compu-tational workload in such computations across parallel processors, achieving sustained near machine-peak per-formance with close-to-ideal load balanced computation-to-processor mapping is inherently difficult. As a result, most of the time, the loads assigned to parallel proces-sors can exhibit significant variations. While there have been numerous past efforts that study this imbalance from the performance viewpoint, to our knowledge, no prior study has considered exploiting the imbalance for re-ducing power consumption during execution. Power con-sumption in large-scale clusters of workstations is be-coming a critical issue as noted by several recent re-search papers from both industry and academia. Focus-ing on sparse matrix computations in which underlying parallel computations and data dependencies can be rep-resented by trees, this paper proposes schemes that save power through voltage/frequency scaling. Our goal is to reduce overall energy consumption by scaling the volt-ages/frequencies of those processors that are not in the critical path; i.e., our approach is oriented towards sav-ing power without incurring performance penalties. 1