Leveraging Task-Parallelism in Energy-Efficient ILU Preconditioners
We analyze the energy-performance balance of a task-parallel computation of an ILU-based preconditioner for the solution of sparse linear systems on multi-core processors. In particular, we elaborate a theoretical model for the power dissipation, and employ it to explore the effect of the processor power states on the time-power-energy interaction for this calculation. Armed with the insights gained from this study, we then introduce two energy-saving mechanisms which, incorporated into the runtime in charge of the parallel execution of the algorithm, improve energy efficiency by 6.9%, with a negligible impact on performance.
KeywordsExecution Time Sparse Linear System Inactive Period Nest Dissection Linear Algebra Operation
Unable to display preview. Download preview PDF.
- 3.AnandTech Forums. Power-consumption scaling with clockspeed and Vcc for the i7-2600K (2011), http://forums.anandtech.com/showthread.php?t=2195927
- 6.Gupta, V., Brett, P., Koufaty, D., Reddy, D., Hahn, S., Schwan, K., Srinivasa, G.: The forgotten ’uncore’: On the energy-efficiency of heterogeneous cores. In: Proc. 2012 USENIX Annual Technical Conference (to appear, 2012)Google Scholar
- 7.HP Corp., Intel Corp., Microsoft Corp., Phoenix Tech. Ltd., and Toshiba Corp. Advanced configuration and power interface specification, revision 5.0 (2011)Google Scholar
- 8.Dongarra, J., et al.: The international ExaScale software project roadmap. Int. J. of High Performance Computing & Applications 25(1), 3–60Google Scholar
- 9.Duranton, M., et al.: The HiPEAC vision (2010), http://www.hipeac.net/roadmap
- 10.Saad, Y.: Iterative Methods for Sparse Linear Systems. SIAM Publications (2003)Google Scholar