Neural networks for pattern recognition
Neural networks for pattern recognition
MAHARADJA: An Embedded System for the Real Time Execution of GRBF Networks
MICRONEURO '99 Proceedings of the 7th International Conference on Microelectronics for Neural, Fuzzy and Bio-Inspired Systems
Accelerating RBF Network Simulation by Using Multimedia Extensions of Modern Microprocessors
ICANN '01 Proceedings of the International Conference on Artificial Neural Networks
Performance Analysis of Intel's MMX and SSE: A Case Study
Euro-Par '01 Proceedings of the 7th International Euro-Par Conference Manchester on Parallel Processing
Hi-index | 0.00 |
In this article we determine the performances of general purpose micro-processors with multimedia extensions, i.e., including a SIMD micro-machine. We studied the Pentium and Pentium II processors with MMX technology to evaluate the capability of the multimedia extensions to simulate in real-time RBF networks using the Mahalanobis distance. For that purpose we implemented the Mahalanobis distance both in MMX assembly code and in C language. Our experiments showed that the Pentium processors with MMX technology are capable of simulating RBF neural networks in real-time with a speed up greater than 3 over usual code in C.