论文信息 - Speed up performances on MIMD machines

Speed up performances on MIMD machines

The authors describe the implementation of a second-order backpropagation algorithm for pattern recognition on a PC, an INMOS transputer network, and Intel hypercubes. The approach was to parallelize a sequential program by identifying the parts of the sequential code that can be executed in parallel and distributing these parts evenly among available processors in the network. The trade-off between communication and computation was investigated. The speedup concurrency is given with respect the number of processors in terms of a sigmoidallike curve.<<ETX>>

[1] D. Roweth,et al. Implementing Neural Network Models on Parallel Computers , 1987, Comput. J..

[2] Jeremy Cook,et al. Parallel neural network simulation using sparse matrix techniques , 1988 .

[3] Kunihiko Fukushima,et al. Neocognitron: A hierarchical neural network capable of visual pattern recognition , 1988, Neural Networks.

[4] Hyunsoo Yoon,et al. A distributed backpropagation algorithm of neural networks on distributed-memory multiprocessors , 1990, [1990 Proceedings] The Third Symposium on the Frontiers of Massively Parallel Computation.

[5] Peter F. Corbett,et al. Communications Overhead and the Expected Speedup of Multidimensional Mesh-Connected Parallel Processors , 1990, J. Parallel Distributed Comput..

[6] Paul Baffes,et al. Design of a neural network simulator on a transputer array , 1987 .