Title :
Parallel neural network training with OpenCL
Author :
Krpan, N. ; Jakobovic, Domagoj
Abstract :
This paper describes the parallelization of neural network training algorithms on heterogeneous architectures with graphical processing units (GPU). The algorithms used for training are particle swarm optimization and backpropagation. Parallel versions of both methods are presented and speedup results are given as compared to the sequential version. The efficiency of parallel training is investigated in regards to various neural network and training parameters.
Keywords :
backpropagation; graphics processing units; neural nets; parallel processing; particle swarm optimisation; OpenCL; backpropagation; heterogeneous architectures; parallel neural network training algorithm; parallel training; parallelization; particle swarm optimization; Backpropagation; Biological neural networks; Graphics processing unit; Memory management; Neurons; Random access memory; Training;
Conference_Titel :
MIPRO, 2012 Proceedings of the 35th International Convention
Conference_Location :
Opatija
Print_ISBN :
978-1-4673-2577-6
