A PARALLEL NEURAL-NETWORK SIMULATOR ON THE CONNECTION MACHINE CM-5

Where present a parallel implementation of artificial neural networks on the connection machine CM-5 and compare it with other parallel impl ementations on SIMD and MIMD architectures. This parallel implementati on was developed with tire goal of efficiency training large neural ne tworks with huge training pattern sets for applications in molecular b iology, in particular the prediction of coding regions in DNA sequence s. The implementation uses training pattern parallelism and makes use of the parallel I/O facilities of the CM-5 and its efficient reduction operations available within the control network to achieve a high sca lability. The parallel simulator obtains a maximum speed of 149.25 MCU PS for training feedforward networks with backpropagation on a 512 pro cessor CM-5 system without using the CM-5 vector facility. The impleme ntation poses no restriction on the type of network topology and works with different batch training algorithms like BP, Quickprop and Rprop .