PaperPerformance prediction of large MIMD systems for parallel neural network simulations☆
References (17)
- et al.
Scientific benchmark characterizations
Parallel Comput.
(1991) - et al.
A massively parallel architecture for a self-organizing neural pattern recognition machine
Computer Vision, Graphics, and Image Processing
(1987) - et al.
Optimal mapping of neural network learning on message-passing multicomputers
J. Parallel Distributed Comput.
(1992) - et al.
Fully connected neural networks: Simulation on massively parallel computers
The Genesis distributed memory benchmarks
Parallel Comput.
(1991)- et al.
Simulating modular neural networks on message-passing multiprocessors
Parallel Comput.
(1991) - et al.
An implementation of back-propagation learning on GF11, a large SIMD parallel computer
Parallel Comput.
(1990) - et al.
Parallelizing the self-organizing feature map on multiprocessor systems
Parallel Comput.
(1991)
Cited by (5)
Parallel codebook design for vector quantization on a message passing MIMD architecture
2002, Parallel ComputingCitation Excerpt :It is not reasonable to compare those results with the ones reported here for many reasons: the platform is different, the parallelization method is not the same, and the speed-up is calculated differently. There are some MIMD parallel implementations of neural networks which have been trained using the competitive learning algorithms such as the Kohonen maps on the GCEL-512 which is a 512-transputer system [20]. In this paper, a master/worker parallel implementation of a VQ algorithm to train a codebook on gray image database has been evaluated using the Alex AVX-2 parallel computer.
Performance of the Alex AVX-2 MIMD architecture in learning the NetTalk database
2004, IEEE Transactions on Neural NetworksPerformance of a backpropagation trained feedforward network on an MIMD architecture
1998, Concurrency Practice and ExperienceOn the implementation of backpropagation on the Alex AVX-2 parallel system
1997, IEEE International Conference on Neural Networks - Conference ProceedingsEstimating the parallel start-up overhead for parallelizing compilers
1997, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- ☆
We would like to acknowledge the University of Amsterdam and Parsytec Gmbh for allowing us to use the GCel located in Amsterdam during the CAMPP '93 programme.