The NAS parallel benchmarks—summary and preliminary results
Proceedings of the 1991 ACM/IEEE conference on Supercomputing
IPDPS '05 Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 4 - Volume 05
Optimizing Compiler for the CELL Processor
Proceedings of the 14th International Conference on Parallel Architectures and Compilation Techniques
The potential of the cell processor for scientific computing
Proceedings of the 3rd conference on Computing frontiers
CellSs: a programming model for the cell BE architecture
Proceedings of the 2006 ACM/IEEE conference on Supercomputing
Dynamic multigrain parallelization on the cell broadband engine
Proceedings of the 12th ACM SIGPLAN symposium on Principles and practice of parallel programming
Implementation and performance analysis of parallel conjugate gradient on the cell broadband engine
IBM Journal of Research and Development
Concurrency and Computation: Practice & Experience
| Hi-index | 0.00 |
The NAS Conjugate Gradient (CG) benchmark is an important scientific kernel used to evaluate machine performance and compare characteristics of different programming models. CG represents a computation and communication paradigm for sparse linear algebra, which is common in scientific fields. In this paper, we present the porting, performance optimization and evaluation of CG on Cell Broadband Engine (CBE). CBE, a heterogeneous multi-core processor with SIMD accelerators, is gaining attention and being deployed on supercomputers and high-end server architectures. We take advantages of CBE's particular architecture to optimize the performance of CG. We also quantify these optimizations and assess their impact. In addition, by exploring distributed nature of CBE, we present trade-off between parallelization and serialization, and Cell-specific data scheduling in its memory hierarchy. Our final result shows that the CG-Cell can achieve more than 4 times speedup over the performance of single comparable PowerPC Processor.