We designed a system for rapidly composing simulations of networks characterized by extreme heterogeneity and scale. Our target domain, computational neuroscience, requires flexibility for rapid and iterative extension and revision of each modeled system. Here we present a graph simulator, which employs one language for defining interoperable models and one for declaring graphs whose vertices are instances of these models. Together these constitute a standard programming model for specifying simulations of complex networks such as those found in neural tissue. Our graphs comprise heterogeneous collections of models whose connections are initialized at runtime for communicating specific model state at specific phases of model execution using efficient collective communication. We demonstrate the applicability of our graph declaration languages for parallel programming, with examples of the 3-dimensional Fourier transform. Finally we show preliminary scaling and performance on the Blue Gene/L supercomputer comparable to that of a stand-alone, optimized implementation of the same algorithm.
By: James Kozloski; Maria Eleftheriou; Blake Fitch; Charles Peck
Published in: RC24811 in 2009
LIMITED DISTRIBUTION NOTICE:
This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.
Questions about this service can be mailed to reports@us.ibm.com .