Title :
Performance of the Alex AVX-2 MIMD architecture in learning the NetTalk database
Author_Institution :
on leave from the Dept. of Comput. & Syst. Eng., Mentor Graphics Corp., Cairo, Egypt
fDate :
3/1/2004 12:00:00 AM
Abstract :
The process of training neural networks on parallel architectures has been used to assess the performance of so many parallel machines. In this paper, we are investigating the implementation of backpropagation (BP) on the Alex AVX-2 coarse-grained MIMD machine. A host-worker parallel implementation is carried out in order to train different networks to learn the NetTalk dictionary. First, a computational model is constructed using a single processor to complete the learning process. Also, a communication model for the host-worker topology is developed in order to compute the communication overhead in the broadcasting/gathering process. Both models are then used to predict the machine performance when p processors are used and a comparison with the actual measured performance of the parallel architecture implementation is carried out. Simulation results show that both models can be used effectively to predict the machine performance for the NetTalk problem. Finally, a comparison between the AVX-2 NetTalk implementation and the performance of other parallel platforms is presented.
Keywords :
backpropagation; feedforward; neural nets; parallel architectures; parallel machines; Alex AVX-2 MIMD architecture; NetTalk database; backpropagation; broadcasting/gathering process; feedforward neural networks; host-worker topology; learning process; multiple instruction multiple data architecture; neural network training; parallel computing; parallel machine; Backpropagation; Broadcasting; Computational modeling; Databases; Dictionaries; Network topology; Neural networks; Parallel architectures; Parallel machines; Predictive models; Artificial Intelligence; Databases, Factual; Neural Networks (Computer);
Journal_Title :
Neural Networks, IEEE Transactions on
DOI :
10.1109/TNN.2004.824274
