Distributed memory matrix-vector multiplication and conjugate gradient algorithms
John Gregg Lewis, Robert A. Geijn · Proceedings - Supercomputing · 1993
The critical bottlenecks in the implementation of the conjugate gradient algorithm on distributed memory computers are the communication requirements of the sparse matrix-vector multiply and of the vector recurrences. The data distribution and communication patterns of five general implementations whose realizations demonstrate that the cost of communication can be overcome to a much larger extent than is often assumed are described. The results also apply to more general settings for matrix-vector products, both sparse and dense.
Cited by 60 works. doi:10.1145/169627.169788
Authors
References
- Solving Problems on Concurrent Processors
- Distributed Routing Algorithms for Broadcasting and Personalized Communication in Hypercubes
- Global combine on mesh architectures with wormhole routing
- Sparse Matrix Computations on Parallel Processor Arrays
- Efficient Communication Primitives on Mesh Architectures with Hardware Routing.
- Optimal Broadcasting in Mesh-Connected Architectures
- Sparse extensions to the FORTRAN Basic Linear Algebra Subprograms
- Matrix-vector multiplication and conjugate gradient algorithms on distributed memory computers
- LAPACK Working Note 29: On Global Combine Operations