Claverol, E.T., Brown, A.D. and Chad, J.E.
Scalable cortical simulations on Beowulf architectures.
Neurocomputing, 43, (1), . (doi:10.1016/S0925-2312(01)00668-3).
Full text not available from this repository.
Biologically motivated simulation of large scale neural networks is a computationally costly task. In this paper, a commodity 8-node Beowulf architecture is proposed as a scalable low cost environment for studies of cortical dynamics, By means of a distributed message-based event-driven framework, the size of memory-limited tractable problems increased 8-fold, resulting in a mere 8.3% increase in elapsed CPU time, attributable to inter-process communication overhead. The attainable network size reached over 106 neu- rons and 2.5 x 108 synapses, with a typical performance of 900 s, Beowulf processing time, per simulated second. (C) 2002 Elsevier Science B.V. All rights reserved.
Actions (login required)