Last updated: April 30, 2009
Introducing PIAF, a program for simulating large-scale, parallel integrate-and-fire neuronal networks. With this innovative software and its easy-to-use graphical user interfaces (GUIs), users can create and simulate neuronal networks efficiently and accurately without having to write code!
Create a network using the editing GUI, simulate, then
load the data using the second GUI, and finally, visualize!
This software is written in C++ with MATLAB frontends. PIAF harnesses its parallel power from PETSc, the Portable Extensible Toolkit for Scientific Computing. All that the user needs to do is download and install PETSc, as explained in the PIAF UserŐs Guide, and then install PIAF.
The interface is intuitive. The methods are efficient. The solutions are accurate.
Neuronal network simulation is now simple.
System Requirements: UNIX / Linux, and it should work on Windows or Mac if run through a terminal, X11, or similar.
UserŐs Guide: download the PIAF UserŐs Guide, current as of 8/30/07
Source Code: download PIAF v1.0 (build 5c) (C++ and MATLAB files)
Updated 1/4/08: fixed bug in parallel data output when no cells monitored on main processor
Updated 11/30/07: fixed some bugs in timestepping methods, better parallel synaptic delivery for timestepping
Updated 11/28/07: better parallel synaptic delivery for quadrature yields much faster execution,
many other small optimizations
(v1.0b4e): Updated 11/20/07: lots of little optimizations, fixed STI data output error,
parallelized EC/DG spike delivery for timestepping
Updated 11/19/07: parallelized EC/DG spike delivery for quadrature, fixed bug to print prop.
Updated 11/17/07: added ability to output EC/DG layer spike data
(v1.0b4): Updated 9/22/07: minor fix to Connectgui
Updated 8/30/07: new queue system for quadrature improves speed!
(v1.0b3): Updated 8/25/07: added Connectgui, for visually creating weight matrices
Updated 8/19/07: adding functionality and fixing bugs in some small special cases
We thank Alexandre Matuszczak for his great creativity in designing the PIAF logo.
We are pleased to acknowledge the generous support of the Sheafor/Lindsay Fund via the ERIT program at Rice's Computer and Information Technology Institute (CITI) and NSF grant DMS-0240058.
Also see the hippos group web page