GeNN: a code generation framework for accelerated brain simulations

Abstract

Large-scale numerical simulations of detailed brain circuit models are important for identifying hypotheses on brain functions and testing their consistency and plausibility. An ongoing challenge for simulating realistic models is, however, computational speed. In this paper, we present the GeNN (GPU-enhanced Neuronal Networks) framework, which aims to facilitate the use of graphics accelerators for computational models of large-scale neuronal networks to address this challenge. GeNN is an open source library that generates code to accelerate the execution of network simulations on NVIDIA GPUs, through a flexible and extensible interface, which does not require in-depth technical knowledge from the users. We present performance benchmarks showing that 200-fold speedup compared to a single core of a CPU can be achieved for a network of one million conductance based Hodgkin-Huxley neurons but that for other models the speedup can differ. GeNN is available for Linux, Mac OS X and Windows platforms. The source code, user manual, tutorials, Wiki, in-depth example projects and all other related information can be found on the project website http://genn-team.github.io/genn/.

Brian 2: an intuitive and efficient neural simulator

“…In addition to the existing standalone mode, it is possible to write plugins for Brian to generate code for these platforms and techniques without modifying the core code, and there are several ongoing projects to do so. These include Brian2GeNN (Stimberg et al, 2018) which uses the GPU-enhanced Neural Network simulator (GeNN; Yavuz et al 2016) to accelerate simulations in some cases by tens to hundreds of times, and Brian2CUDA…”

Reproducibility in Computational Neuroscience Models and Simulations

“…Brian then uses Python commands to construct groups of such neurons and connect them. Numerical integration in Brian is performed using C++ or graphics processing unit (GPU)-based generated code via GeNN [14], [79]. The latter strategy is advantageous because even low end GPUs typically support many more parallel calculation pipelines than CPUs, thereby offering an inexpensive route to high speed simulations.…”

“…Largescale network simulations may require High Performance Computers (HPCs), or Graphical Processing Units (GPUs) [13], [14]. Still more problematic are simulations that use highly specialized systems such as neuromorphic chips [15], [16], or SpiNNaker chips [17], [18], hardwares that are not commercially available.…”

Quantization of Map-Based Neuronal Model for Embedded Simulations of Neurobiological Networks in Real-Time

“…They mainly rely on the development and implementation of a spiking neuron model and synapses that replicate the dynamics of neurons as close to real biological neurons as possible that fits in computational hardware. Two main directions of these studies are the use of analog computing simulating the network with specially designed VLSI integrated circuits (Indiveri et al, 2006;Silver et al, 2007) and numerical simulation of the model with specially designed microchips, GPUs FPGAs or DSPs (Yavuz et al, 2016;Zbrzeski et al, 2016;Cheung et al, 2016;Brette et al, 2007 and references therein). This paper focuses on the design, implementation and validation of a neuronal model that could be used in real-time simulations of relatively large networks using off-the-shelf 32-bit microprocessors.…”