Leveraging task-parallelism in energy-efficient ILU preconditioners

Authors:
José I. Aliaga;Manuel F. Dolz;Alberto F. Martín;Rafael Mayo;Enrique S. Quintana-Ortí
Affiliations:
Dpto. de Ingeniería y Ciencia de Computadores, Universitat Jaume I, Castellón, Spain;Dpto. de Ingeniería y Ciencia de Computadores, Universitat Jaume I, Castellón, Spain;Centre Internacional de Mètodes Numèrics en Enginyeria (CIMNE), Castelldefels, Spain;Dpto. de Ingeniería y Ciencia de Computadores, Universitat Jaume I, Castellón, Spain;Dpto. de Ingeniería y Ciencia de Computadores, Universitat Jaume I, Castellón, Spain
Venue:
ICT-GLOW'12 Proceedings of the Second international conference on ICT as Key Technology against Global Warming
Year:
2012

Citing 7
Cited 0

Iterative Methods for Sparse Linear Systems

Iterative Methods for Sparse Linear Systems
Analyzing the Energy-Time Trade-Off in High-Performance Computing Applications

IEEE Transactions on Parallel and Distributed Systems
Green Supercomputing Comes of Age

IT Professional
Energy-efficient algorithms

Communications of the ACM
The International Exascale Software Project roadmap

International Journal of High Performance Computing Applications
Exploiting thread-level parallelism in the iterative solution of sparse linear systems

Parallel Computing
The forgotten 'uncore': on the energy-efficiency of heterogeneous cores

USENIX ATC'12 Proceedings of the 2012 USENIX conference on Annual Technical Conference

Quantified Score

Hi-index	0.00

Visualization

Abstract

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.