Éditeur(s) : HAL CCSD
Résumé : 6 pages
In a Spiking Neural Networks (SNN), spike emissions are sparsely and irregularly distributed both in time and in the network architecture. Since a current feature of SNNs is a low average activity, efficient implementations of SNNs are usually based on an Event-Driven Simulation (EDS). On the other hand, simulations of large scale neural networks can take advantage of distributing the neurons on a set of processors (either workstation cluster or parallel computer). This article presents DAMNED, a large scale SNN simulation framework able to gather the benefits of EDS and parallel computing. Two levels of parallelism are combined: Distributed mapping of the neural topology, at the network level, and local multithreaded allocation of resources for simultaneous processing of events, at the neuron level. Based on the causality of events, a distributed solution is proposed for solving the complex problem of scheduling without synchronization barrier.
https://hal.archives-ouvertes.fr/hal-00015137

hal-00015137
https://hal.archives-ouvertes.fr/hal-00015137
https://hal.archives-ouvertes.fr/hal-00015137v2/document
https://hal.archives-ouvertes.fr/hal-00015137/file/hal_damned.pdf
ARXIV : cs.NE/0512018

Éditeur(s) : HAL CCSD Geological Society of America
Résumé : International audience
Lakes are major depositional systems for which the related depositional processes have long been considered relatively simple. Breaking this statement, this study presents a detailed analysis of deposits in Lake Saint-Jean, the third largest natural lake in Québec. In addition to postglacial deltaic and coastal depositional systems fringing the lake, current-controlled features such as a large subaqueous prograding wedge and three sediment drifts have been identifi ed in its central portion based on two-dimensional (2-D) acoustic high-resolution subbottom profiles. The large subaqueous prograding wedge is a 4-km-long and up to 15-m-thick heterolithic shelf-like construction in the southeastern part of the lake. The three sediment drifts are 0.1–0.5-km-long and 2–5-mthick mud mounds distributed on the lake floor in the central portion of the lake. Diatom analyses and radiocarbon dating show that the development of these current-controlled features occurred during the lacustrine phase, after the disconnection with the postglacial marine Laflamme Gulf at 8.5 cal. k.y. B.P. Depositional facies show evidence of recurrent bottom-current activity. Related deposits alternate with pelagic sedimentation stages characterized by the settling of mud and biogenic accumulations. We investigated the origin of bottom currents using a numerical simulation (SYMPHONIE, an oceanographic model), with the aim of modeling wind-induced lake-scale water circulation. Simulations suggest that the subaqueous prograding wedge and the three sediment drifts result from wind-induced bottom currents generated by storm events having wind speed greater than 10 m s–1. Such strong winds are able to significantly affect sedimentation in the central portion of Lake Saint-Jean. Theresulting wind-induced sedimentary features were integrated into a refi ned lacustrine depositional model that summarizes the evolution of a group of water bodies referred to as “wind-driven water bodies.” This study applies a new tool for lake strata characterization and highlights the potential diffi cultyin differentiating them from marine deposits in the geological record.
ISSN: 0016-7606

insu-01184724
https://hal-insu.archives-ouvertes.fr/insu-01184724
DOI : 10.1130/B31145.1

Éditeur(s) : HAL CCSD Springer
Résumé : International audience
This paper presents a spiking neural network simulator suitablefor biologically plausible large neural networks, named DAMNEDfor “Distributed And Multi-threaded Neural Event-Driven”. The simulatoris designed to run efficiently on a variety of hardware. DAMNEDmakes use of multi-threaded programming and non-blocking communicationsin order to optimize communications and computations overlap.This paper details the even-driven architecture of the simulator. Someoriginal contributions are presented, such as the handling of a distributedvirtual clock and an efficient circular event queue taking into accountspike propagation delays. DAMNED is evaluated on a cluster of computersfor networks from 103 to 105 neurons. Simulation and networkcreation speedups are presented. Finally, scalability is discussed regardingnumber of processors, network size and activity of the simulated NN.
Engineering Applications of Neural Networks: 11th International Conference, EANN 2009
London, United Kingdom

hal-01437753
https://hal.archives-ouvertes.fr/hal-01437753
DOI : 10.1007/978-3-642-03969-0_33