Performance analysis of asynchronous Jacobi's method implemented in MPI, SHMEM and OpenMP

Authors:
Iain Bethune;J Mark Bull;Nicholas J Dingle;Nicholas J Higham
Affiliations:
EPCC, The University of Edinburgh, UK;EPCC, The University of Edinburgh, UK;School of Mathematics, University of Manchester, UK;School of Mathematics, University of Manchester, UK
Venue:
International Journal of High Performance Computing Applications
Year:
2014

Citing 11
Cited 0

Parallel and distributed computation: numerical methods

Parallel and distributed computation: numerical methods
Asynchronous Iterative Methods for Multiprocessors

Journal of the ACM (JACM)
On asynchronous iterations

Journal of Computational and Applied Mathematics - Special issue on numerical analysis 2000 Vol. III: linear algebra
Numerical Performance of an Asynchronous Jacobi Iteration

CONPAR '92/ VAPP V Proceedings of the Second Joint International Conference on Vector and Parallel Processing: Parallel Processing
Iterative Methods for Sparse Linear Systems

Iterative Methods for Sparse Linear Systems
Coupling Dynamic Load Balancing with Asynchronism in Iterative Algorithms on the Computational Grid

IPDPS '03 Proceedings of the 17th International Symposium on Parallel and Distributed Processing
Performance Comparison of Parallel Programming Environments for Implementing AIAC Algorithms

The Journal of Supercomputing
Topology mapping for Blue Gene/L supercomputer

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
Extracting state-based performance metrics using asynchronous iterative techniques

Performance Evaluation
Communication-avoiding krylov subspace methods

Communication-avoiding krylov subspace methods
Introducing OpenSHMEM: SHMEM for the PGAS community

Proceedings of the Fourth Conference on Partitioned Global Address Space Programming Model

Quantified Score

Hi-index	0.00

Visualization

Abstract

Ever-increasing core counts create the need to develop parallel algorithms that avoid closely coupled execution across all cores. We present performance analysis of several parallel asynchronous implementations of Jacobi's method for solving systems of linear equations, using MPI, SHMEM and OpenMP. In particular we have solved systems of over 4 billion unknowns using up to 32,768 processes on a Cray XE6 supercomputer. We show that the precise implementation details of asynchronous algorithms can strongly affect the resulting performance and convergence behaviour of our solvers in unexpected ways, discuss how our specific implementations could be generalised to other classes of problem, and suggest how existing parallel programming models might be extended to allow asynchronous algorithms to be expressed more easily.