Abstract
Purchase PDF (205 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (0)
Purchase PDF (77 K)
doi:10.1016/j.parco.2007.06.006 
Copyright © 2007 Elsevier B.V. All rights reserved.
Optimizing a conjugate gradient solver with non-blocking collective operations
Received 22 December 2006;
References and further reading may be available for this article. To view references and further reading you must purchase this article.
Abstract
This paper presents a case study that analyzes the suitability and usage of non-blocking collective operations in parallel applications. As with their point-to-point counterparts, non-blocking collective operations provide the ability to overlap communication with computation and to avoid unnecessary synchronization. These operations are provided for MPI programs with LibNBC, a portable low-overhead implementation of non-blocking collective operations built on MPI-1. The straightforward applicability of the LibNBC is demonstrated by incorporating non-blocking collective operations into a parallel conjugate gradient solver. Although only minor changes are required to use them, non-blocking collective operations allow most of the communication costs to be hidden and provide performance improvements of up to 34%. We also show that, because of overlap, there is no significant performance difference between Gigabit Ethernet and InfiniBandTM for special cases of our calculation.
Keywords: Message passing interface (MPI); Communication; Computation overlap; Collective operations; Non-blocking collective operations; Poisson solver
Article Outline
- 1. Introduction
- 1.1. Related work
- 2. Implementing non-blocking collective operations
- 2.1. The scheduling engine
- 2.2. Building a schedule
- 2.3. Schedule execution
- 3. Optimization of linear solvers
- 3.1. Case study: three-dimensional Poisson equation
- 3.2. Domain decomposition
- 3.3. Design and optimization of the CG solver
- 3.4. Parallel matrix vector product
- 3.5. Benchmark results
- 3.6. Comparison to non-blocking point-to-point messaging
- 3.7. Optimization impact on other linear solvers
- 4. Conclusions and future work
- Acknowledgements
- References
