Meeting the memory challenges of brain-scale network simulation

BrainPy, a flexible, integrative, efficient, and extensible framework for general-purpose brain dynamics programming

Elucidating the intricate neural mechanisms underlying brain functions requires integrative brain dynamics modeling. To facilitate this process, it is crucial to develop a general-purpose programming framework that allows users to freely define neural models across multiple scales, efficiently simulate, train, and analyze model dynamics, and conveniently incorporate new modeling approaches. In response to this need, we present BrainPy. BrainPy leverages the advanced just-in-time (JIT) compilation capabilities of JAX and XLA to provide a powerful infrastructure tailored for brain dynamics programming. It offers an integrated platform for building, simulating, training, and analyzing brain dynamics models. Models defined in BrainPy can be JIT compiled into binary instructions for various devices, including Central Processing Unit, Graphics Processing Unit, and Tensor Processing Unit, which ensures high-running performance comparable to native C or CUDA. Additionally, BrainPy features an extensible architecture that allows for easy expansion of new infrastructure, utilities, and machine-learning approaches. This flexibility enables researchers to incorporate cutting-edge techniques and adapt the framework to their specific needs.

Efficient parameter calibration and real-time simulation of large-scale spiking neural networks with GeNN and NEST

Spiking neural networks (SNNs) represent the state-of-the-art approach to the biologically realistic modeling of nervous system function. The systematic calibration for multiple free model parameters is necessary to achieve robust network function and demands high computing power and large memory resources. Special requirements arise from closed-loop model simulation in virtual environments and from real-time simulation in robotic application. Here, we compare two complementary approaches to efficient large-scale and real-time SNN simulation. The widely used NEural Simulation Tool (NEST) parallelizes simulation across multiple CPU cores. The GPU-enhanced Neural Network (GeNN) simulator uses the highly parallel GPU-based architecture to gain simulation speed. We quantify fixed and variable simulation costs on single machines with different hardware configurations. As a benchmark model, we use a spiking cortical attractor network with a topology of densely connected excitatory and inhibitory neuron clusters with homogeneous or distributed synaptic time constants and in comparison to the random balanced network. We show that simulation time scales linearly with the simulated biological model time and, for large networks, approximately linearly with the model size as dominated by the number of synaptic connections. Additional fixed costs with GeNN are almost independent of model size, while fixed costs with NEST increase linearly with model size. We demonstrate how GeNN can be used for simulating networks with up to 3.5 · 106 neurons (> 3 · 1012synapses) on a high-end GPU, and up to 250, 000 neurons (25 · 109 synapses) on a low-cost GPU. Real-time simulation was achieved for networks with 100, 000 neurons. Network calibration and parameter grid search can be efficiently achieved using batch processing. We discuss the advantages and disadvantages of both approaches for different use cases.

A Modular Workflow for Performance Benchmarking of Neuronal Network Simulations

Modern computational neuroscience strives to develop complex network models to explain dynamics and function of brains in health and disease. This process goes hand in hand with advancements in the theory of neuronal networks and increasing availability of detailed anatomical data on brain connectivity. Large-scale models that study interactions between multiple brain areas with intricate connectivity and investigate phenomena on long time scales such as system-level learning require progress in simulation speed. The corresponding development of state-of-the-art simulation engines relies on information provided by benchmark simulations which assess the time-to-solution for scientifically relevant, complementary network models using various combinations of hardware and software revisions. However, maintaining comparability of benchmark results is difficult due to a lack of standardized specifications for measuring the scaling performance of simulators on high-performance computing (HPC) systems. Motivated by the challenging complexity of benchmarking, we define a generic workflow that decomposes the endeavor into unique segments consisting of separate modules. As a reference implementation for the conceptual workflow, we develop beNNch: an open-source software framework for the configuration, execution, and analysis of benchmarks for neuronal network simulations. The framework records benchmarking data and metadata in a unified way to foster reproducibility. For illustration, we measure the performance of various versions of the NEST simulator across network models with different levels of complexity on a contemporary HPC system, demonstrating how performance bottlenecks can be identified, ultimately guiding the development toward more efficient simulation technology.

Routing Brain Traffic Through the Von Neumann Bottleneck: Parallel Sorting and Refactoring

Generic simulation code for spiking neuronal networks spends the major part of the time in the phase where spikes have arrived at a compute node and need to be delivered to their target neurons. These spikes were emitted over the last interval between communication steps by source neurons distributed across many compute nodes and are inherently irregular and unsorted with respect to their targets. For finding those targets, the spikes need to be dispatched to a three-dimensional data structure with decisions on target thread and synapse type to be made on the way. With growing network size, a compute node receives spikes from an increasing number of different source neurons until in the limit each synapse on the compute node has a unique source. Here, we show analytically how this sparsity emerges over the practically relevant range of network sizes from a hundred thousand to a billion neurons. By profiling a production code we investigate opportunities for algorithmic changes to avoid indirections and branching. Every thread hosts an equal share of the neurons on a compute node. In the original algorithm, all threads search through all spikes to pick out the relevant ones. With increasing network size, the fraction of hits remains invariant but the absolute number of rejections grows. Our new alternative algorithm equally divides the spikes among the threads and immediately sorts them in parallel according to target thread and synapse type. After this, every thread completes delivery solely of the section of spikes for its own neurons. Independent of the number of threads, all spikes are looked at only two times. The new algorithm halves the number of instructions in spike delivery which leads to a reduction of simulation time of up to 40 %. Thus, spike delivery is a fully parallelizable process with a single synchronization point and thereby well suited for many-core systems. Our analysis indicates that further progress requires a reduction of the latency that the instructions experience in accessing memory. The study provides the foundation for the exploration of methods of latency hiding like software pipelining and software-induced prefetching.

Dynamical Characteristics of Recurrent Neuronal Networks Are Robust Against Low Synaptic Weight Resolution

The representation of the natural-density, heterogeneous connectivity of neuronal network models at relevant spatial scales remains a challenge for Computational Neuroscience and Neuromorphic Computing. In particular, the memory demands imposed by the vast number of synapses in brain-scale network simulations constitute a major obstacle. Limiting the number resolution of synaptic weights appears to be a natural strategy to reduce memory and compute load. In this study, we investigate the effects of a limited synaptic-weight resolution on the dynamics of recurrent spiking neuronal networks resembling local cortical circuits and develop strategies for minimizing deviations from the dynamics of networks with high-resolution synaptic weights. We mimic the effect of a limited synaptic weight resolution by replacing normally distributed synaptic weights with weights drawn from a discrete distribution, and compare the resulting statistics characterizing firing rates, spike-train irregularity, and correlation coefficients with the reference solution. We show that a naive discretization of synaptic weights generally leads to a distortion of the spike-train statistics. If the weights are discretized such that the mean and the variance of the total synaptic input currents are preserved, the firing statistics remain unaffected for the types of networks considered in this study. For networks with sufficiently heterogeneous in-degrees, the firing statistics can be preserved even if all synaptic weights are replaced by the mean of the weight distribution. We conclude that even for simple networks with non-plastic neurons and synapses, a discretization of synaptic weights can lead to substantial deviations in the firing statistics unless the discretization is performed with care and guided by a rigorous validation process. For the network model used in this study, the synaptic weights can be replaced by low-resolution weights without affecting its macroscopic dynamical characteristics, thereby saving substantial amounts of memory.

Towards simulations of long-term behavior of neural networks: Modeling synaptic plasticity of connections within and between human brain regions

Simulations of neural networks can be used to study the direct effect of internal or external changes on brain dynamics. However, some changes are not immediate but occur on the timescale of weeks, months, or years. Examples include effects of strokes, surgical tissue removal, or traumatic brain injury but also gradual changes during brain development. Simulating network activity over a long time, even for a small number of nodes, is a computational challenge. Here, we model a coupled network of human brain regions with a modified Wilson-Cowan model representing dynamics for each region and with synaptic plasticity adjusting connection weights within and between regions. Using strategies ranging from different models for plasticity, vectorization and a different differential equation solver setup, we achieved one second runtime for one second biological time.

Simulation of Large Scale Neural Models With Event-Driven Connectivity Generation

Accurate simulations of brain structures is a major problem in neuroscience. Many works are dedicated to design better models or to develop more efficient simulation schemes. In this paper, we propose a hybrid simulation scheme that combines time-stepping second-order integration of Hodgkin-Huxley (HH) type neurons with event-driven updating of the synaptic currents. As the HH model is a continuous model, there is no explicit spike events. Thus, in order to preserve the accuracy of the integration method, a spike detection algorithm is developed that accurately determines spike times. This approach allows us to regenerate the outgoing connections at each event, thereby avoiding the storage of the connectivity. Consequently, memory consumption is significantly reduced while preserving execution time and accuracy of the simulations, especially the spike times of detailed point neuron models. The efficiency of the method, implemented in the SiReNe software, is demonstrated by the simulation of a striatum model which consists of more than 106 neurons and 108 synapses (each neuron has a fan-out of 504 post-synaptic neurons), under normal and Parkinson's conditions.

Efficient Communication in Distributed Simulations of Spiking Neuronal Networks With Gap Junctions

Investigating the dynamics and function of large-scale spiking neuronal networks with realistic numbers of synapses is made possible today by state-of-the-art simulation code that scales to the largest contemporary supercomputers. However, simulations that involve electrical interactions, also called gap junctions, besides chemical synapses scale only poorly due to a communication scheme that collects global data on each compute node. In comparison to chemical synapses, gap junctions are far less abundant. To improve scalability we exploit this sparsity by integrating an existing framework for continuous interactions with a recently proposed directed communication scheme for spikes. Using a reference implementation in the NEST simulator we demonstrate excellent scalability of the integrated framework, accelerating large-scale simulations with gap junctions by more than an order of magnitude. This allows, for the first time, the efficient exploration of the interactions of chemical and electrical coupling in large-scale neuronal networks models with natural synapse density distributed across thousands of compute nodes.

Computational geometry for modeling neural populations: From visualization to simulation

The importance of a mesoscopic description level of the brain has now been well established. Rate based models are widely used, but have limitations. Recently, several extremely efficient population-level methods have been proposed that go beyond the characterization of a population in terms of a single variable. Here, we present a method for simulating neural populations based on two dimensional (2D) point spiking neuron models that defines the state of the population in terms of a density function over the neural state space. Our method differs in that we do not make the diffusion approximation, nor do we reduce the state space to a single dimension (1D). We do not hard code the neural model, but read in a grid describing its state space in the relevant simulation region. Novel models can be studied without even recompiling the code. The method is highly modular: variations of the deterministic neural dynamics and the stochastic process can be investigated independently. Currently, there is a trend to reduce complex high dimensional neuron models to 2D ones as they offer a rich dynamical repertoire that is not available in 1D, such as limit cycles. We will demonstrate that our method is ideally suited to investigate noise in such systems, replicating results obtained in the diffusion limit and generalizing them to a regime of large jumps. The joint probability density function is much more informative than 1D marginals, and we will argue that the study of 2D systems subject to noise is important complementary to 1D systems.

Extremely Scalable Spiking Neuronal Network Simulation Code: From Laptops to Exascale Computers

State-of-the-art software tools for neuronal network simulations scale to the largest computing systems available today and enable investigations of large-scale networks of up to 10 % of the human cortex at a resolution of individual neurons and synapses. Due to an upper limit on the number of incoming connections of a single neuron, network connectivity becomes extremely sparse at this scale. To manage computational costs, simulation software ultimately targeting the brain scale needs to fully exploit this sparsity. Here we present a two-tier connection infrastructure and a framework for directed communication among compute nodes accounting for the sparsity of brain-scale networks. We demonstrate the feasibility of this approach by implementing the technology in the NEST simulation code and we investigate its performance in different scaling scenarios of typical network simulations. Our results show that the new data structures and communication scheme prepare the simulation kernel for post-petascale high-performance computing facilities without sacrificing performance in smaller systems.

The NEST Dry-Run Mode: Efficient Dynamic Analysis of Neuronal Network Simulation Code

NEST is a simulator for spiking neuronal networks that commits to a general purpose approach: It allows for high flexibility in the design of network models, and its applications range from small-scale simulations on laptops to brain-scale simulations on supercomputers. Hence, developers need to test their code for various use cases and ensure that changes to code do not impair scalability. However, running a full set of benchmarks on a supercomputer takes up precious compute-time resources and can entail long queuing times. Here, we present the NEST dry-run mode, which enables comprehensive dynamic code analysis without requiring access to high-performance computing facilities. A dry-run simulation is carried out by a single process, which performs all simulation steps except communication as if it was part of a parallel environment with many processes. We show that measurements of memory usage and runtime of neuronal network simulations closely match the corresponding dry-run data. Furthermore, we demonstrate the successful application of the dry-run mode in the areas of profiling and performance modeling.

Population density equations for stochastic processes with memory kernels

We present a method for solving population density equations (PDEs)--a mean-field technique describing homogeneous populations of uncoupled neurons-where the populations can be subject to non-Markov noise for arbitrary distributions of jump sizes. The method combines recent developments in two different disciplines that traditionally have had limited interaction: computational neuroscience and the theory of random networks. The method uses a geometric binning scheme, based on the method of characteristics, to capture the deterministic neurodynamics of the population, separating the deterministic and stochastic process cleanly. We can independently vary the choice of the deterministic model and the model for the stochastic process, leading to a highly modular numerical solution strategy. We demonstrate this by replacing the master equation implicit in many formulations of the PDE formalism by a generalization called the generalized Montroll-Weiss equation-a recent result from random network theory-describing a random walker subject to transitions realized by a non-Markovian process. We demonstrate the method for leaky- and quadratic-integrate and fire neurons subject to spike trains with Poisson and gamma-distributed interspike intervals. We are able to model jump responses for both models accurately to both excitatory and inhibitory input under the assumption that all inputs are generated by one renewal process.

Constructing Neuronal Network Models in Massively Parallel Environments

Recent advances in the development of data structures to represent spiking neuron network models enable us to exploit the complete memory of petascale computers for a single brain-scale network simulation. In this work, we investigate how well we can exploit the computing power of such supercomputers for the creation of neuronal networks. Using an established benchmark, we divide the runtime of simulation code into the phase of network construction and the phase during which the dynamical state is advanced in time. We find that on multi-core compute nodes network creation scales well with process-parallel code but exhibits a prohibitively large memory consumption. Thread-parallel network creation, in contrast, exhibits speedup only up to a small number of threads but has little overhead in terms of memory. We further observe that the algorithms creating instances of model neurons and their connections scale well for networks of ten thousand neurons, but do not show the same speedup for networks of millions of neurons. Our work uncovers that the lack of scaling of thread-parallel network creation is due to inadequate memory allocation strategies and demonstrates that thread-optimized memory allocators recover excellent scaling. An analysis of the loop order used for network construction reveals that more complex tests on the locality of operations significantly improve scaling and reduce runtime by allowing construction algorithms to step through large networks more efficiently than in existing code. The combination of these techniques increases performance by an order of magnitude and harnesses the increasingly parallel compute power of the compute nodes in high-performance clusters and supercomputers.

Synapse-Centric Mapping of Cortical Models to the SpiNNaker Neuromorphic Architecture

While the adult human brain has approximately 8.8 × 10(10) neurons, this number is dwarfed by its 1 × 10(15) synapses. From the point of view of neuromorphic engineering and neural simulation in general this makes the simulation of these synapses a particularly complex problem. SpiNNaker is a digital, neuromorphic architecture designed for simulating large-scale spiking neural networks at speeds close to biological real-time. Current solutions for simulating spiking neural networks on SpiNNaker are heavily inspired by work on distributed high-performance computing. However, while SpiNNaker shares many characteristics with such distributed systems, its component nodes have much more limited resources and, as the system lacks global synchronization, the computation performed on each node must complete within a fixed time step. We first analyze the performance of the current SpiNNaker neural simulation software and identify several problems that occur when it is used to simulate networks of the type often used to model the cortex which contain large numbers of sparsely connected synapses. We then present a new, more flexible approach for mapping the simulation of such networks to SpiNNaker which solves many of these problems. Finally we analyze the performance of our new approach using both benchmarks, designed to represent cortical connectivity, and larger, functional cortical models. In a benchmark network where neurons receive input from 8000 STDP synapses, our new approach allows 4× more neurons to be simulated on each SpiNNaker core than has been previously possible. We also demonstrate that the largest plastic neural network previously simulated on neuromorphic hardware can be run in real time using our new approach: double the speed that was previously achieved. Additionally this network contains two types of plastic synapse which previously had to be trained separately but, using our new approach, can be trained simultaneously.

Optimizing NEURON Simulation Environment Using Remote Memory Access with Recursive Doubling on Distributed Memory Systems

Increase in complexity of neuronal network models escalated the efforts to make NEURON simulation environment efficient. The computational neuroscientists divided the equations into subnets amongst multiple processors for achieving better hardware performance. On parallel machines for neuronal networks, interprocessor spikes exchange consumes large section of overall simulation time. In NEURON for communication between processors Message Passing Interface (MPI) is used. MPI_Allgather collective is exercised for spikes exchange after each interval across distributed memory systems. The increase in number of processors though results in achieving concurrency and better performance but it inversely affects MPI_Allgather which increases communication time between processors. This necessitates improving communication methodology to decrease the spikes exchange time over distributed memory systems. This work has improved MPI_Allgather method using Remote Memory Access (RMA) by moving two-sided communication to one-sided communication, and use of recursive doubling mechanism facilitates achieving efficient communication between the processors in precise steps. This approach enhanced communication concurrency and has improved overall runtime making NEURON more efficient for simulation of large neuronal network models.

Large-Scale Simulations of Plastic Neural Networks on Neuromorphic Hardware

SpiNNaker is a digital, neuromorphic architecture designed for simulating large-scale spiking neural networks at speeds close to biological real-time. Rather than using bespoke analog or digital hardware, the basic computational unit of a SpiNNaker system is a general-purpose ARM processor, allowing it to be programmed to simulate a wide variety of neuron and synapse models. This flexibility is particularly valuable in the study of biological plasticity phenomena. A recently proposed learning rule based on the Bayesian Confidence Propagation Neural Network (BCPNN) paradigm offers a generic framework for modeling the interaction of different plasticity mechanisms using spiking neurons. However, it can be computationally expensive to simulate large networks with BCPNN learning since it requires multiple state variables for each synapse, each of which needs to be updated every simulation time-step. We discuss the trade-offs in efficiency and accuracy involved in developing an event-based BCPNN implementation for SpiNNaker based on an analytical solution to the BCPNN equations, and detail the steps taken to fit this within the limited computational and memory resources of the SpiNNaker architecture. We demonstrate this learning rule by learning temporal sequences of neural activity within a recurrent attractor network which we simulate at scales of up to 2.0 × 104 neurons and 5.1 × 107 plastic synapses: the largest plastic neural network ever to be simulated on neuromorphic hardware. We also run a comparable simulation on a Cray XC-30 supercomputer system and find that, if it is to match the run-time of our SpiNNaker simulation, the super computer system uses approximately 45× more power. This suggests that cheaper, more power efficient neuromorphic systems are becoming useful discovery tools in the study of plasticity in large-scale brain models.

A unified framework for spiking and gap-junction interactions in distributed neuronal network simulations

Contemporary simulators for networks of point and few-compartment model neurons come with a plethora of ready-to-use neuron and synapse models and support complex network topologies. Recent technological advancements have broadened the spectrum of application further to the efficient simulation of brain-scale networks on supercomputers. In distributed network simulations the amount of spike data that accrues per millisecond and process is typically low, such that a common optimization strategy is to communicate spikes at relatively long intervals, where the upper limit is given by the shortest synaptic transmission delay in the network. This approach is well-suited for simulations that employ only chemical synapses but it has so far impeded the incorporation of gap-junction models, which require instantaneous neuronal interactions. Here, we present a numerical algorithm based on a waveform-relaxation technique which allows for network simulations with gap junctions in a way that is compatible with the delayed communication strategy. Using a reference implementation in the NEST simulator, we demonstrate that the algorithm and the required data structures can be smoothly integrated with existing code such that they complement the infrastructure for spiking connections. To show that the unified framework for gap-junction and spiking interactions achieves high performance and delivers high accuracy in the presence of gap junctions, we present benchmarks for workstations, clusters, and supercomputers. Finally, we discuss limitations of the novel technology.

Spiking network simulation code for petascale computers

Brain-scale networks exhibit a breathtaking heterogeneity in the dynamical properties and parameters of their constituents. At cellular resolution, the entities of theory are neurons and synapses and over the past decade researchers have learned to manage the heterogeneity of neurons and synapses with efficient data structures. Already early parallel simulation codes stored synapses in a distributed fashion such that a synapse solely consumes memory on the compute node harboring the target neuron. As petaflop computers with some 100,000 nodes become increasingly available for neuroscience, new challenges arise for neuronal network simulation software: Each neuron contacts on the order of 10,000 other neurons and thus has targets only on a fraction of all compute nodes; furthermore, for any given source neuron, at most a single synapse is typically created on any compute node. From the viewpoint of an individual compute node, the heterogeneity in the synaptic target lists thus collapses along two dimensions: the dimension of the types of synapses and the dimension of the number of synapses of a given type. Here we present a data structure taking advantage of this double collapse using metaprogramming techniques. After introducing the relevant scaling scenario for brain-scale simulations, we quantitatively discuss the performance on two supercomputers. We show that the novel architecture scales to the largest petascale supercomputers available today.

CyNEST: a maintainable Cython-based interface for the NEST simulator

NEST is a simulator for large-scale networks of spiking point neuron models (Gewaltig and Diesmann, 2007). Originally, simulations were controlled via the Simulation Language Interpreter (SLI), a built-in scripting facility implementing a language derived from PostScript (Adobe Systems, Inc., 1999). The introduction of PyNEST (Eppler et al., 2008), the Python interface for NEST, enabled users to control simulations using Python. As the majority of NEST users found PyNEST easier to use and to combine with other applications, it immediately displaced SLI as the default NEST interface. However, developing and maintaining PyNEST has become increasingly difficult over time. This is partly because adding new features requires writing low-level C++ code intermixed with calls to the Python/C API, which is unrewarding. Moreover, the Python/C API evolves with each new version of Python, which results in a proliferation of version-dependent code branches. In this contribution we present the re-implementation of PyNEST in the Cython language, a superset of Python that additionally supports the declaration of C/C++ types for variables and class attributes, and provides a convenient foreign function interface (FFI) for invoking C/C++ routines (Behnel et al., 2011). Code generation via Cython allows the production of smaller and more maintainable bindings, including increased compatibility with all supported Python releases without additional burden for NEST developers. Furthermore, this novel approach opens up the possibility to support alternative implementations of the Python language at no cost given a functional Cython back-end for the corresponding implementation, and also enables cross-compilation of Python bindings for embedded systems and supercomputers alike.

HRLSim: a high performance spiking neural network simulator for GPGPU clusters

Modeling of large-scale spiking neural models is an important tool in the quest to understand brain function and subsequently create real-world applications. This paper describes a spiking neural network simulator environment called HRL Spiking Simulator (HRLSim). This simulator is suitable for implementation on a cluster of general purpose graphical processing units (GPGPUs). Novel aspects of HRLSim are described and an analysis of its performance is provided for various configurations of the cluster. With the advent of inexpensive GPGPU cards and compute power, HRLSim offers an affordable and scalable tool for design, real-time simulation, and analysis of large-scale spiking neural networks.

Increasing quality and managing complexity in neuroinformatics software development with continuous integration

High quality neuroscience research requires accurate, reliable and well maintained neuroinformatics applications. As software projects become larger, offering more functionality and developing a denser web of interdependence between their component parts, we need more sophisticated methods to manage their complexity. If complexity is allowed to get out of hand, either the quality of the software or the speed of development suffer, and in many cases both. To address this issue, here we develop a scalable, low-cost and open source solution for continuous integration (CI), a technique which ensures the quality of changes to the code base during the development procedure, rather than relying on a pre-release integration phase. We demonstrate that a CI-based workflow, due to rapid feedback about code integration problems and tracking of code health measures, enabled substantial increases in productivity for a major neuroinformatics project and additional benefits for three further projects. Beyond the scope of the current study, we identify multiple areas in which CI can be employed to further increase the quality of neuroinformatics projects by improving development practices and incorporating appropriate development tools. Finally, we discuss what measures can be taken to lower the barrier for developers of neuroinformatics applications to adopt this useful technique.

Supercomputers ready for use as discovery machines for neuroscience

NEST is a widely used tool to simulate biological spiking neural networks. Here we explain the improvements, guided by a mathematical model of memory consumption, that enable us to exploit for the first time the computational power of the K supercomputer for neuroscience. Multi-threaded components for wiring and simulation combine 8 cores per MPI process to achieve excellent scaling. K is capable of simulating networks corresponding to a brain area with 10⁸ neurons and 10¹² synapses in the worst case scenario of random connectivity; for larger networks of the brain its hierarchical organization can be exploited to constrain the number of communicating computer nodes. We discuss the limits of the software technology, comparing maximum filling scaling plots for K and the JUGENE BG/P system. The usability of these machines for network simulations has become comparable to running simulations on a single PC. Turn-around times in the range of minutes even for the largest systems enable a quasi interactive working style and render simulations on this scale a practical tool for computational neuroscience.