Back...

Erasmus MC and Maxeler join forces to make brain simulation easy and affordable

The Neuroscience department of the Erasmus Medical Center (Netherlands) and Maxeler Technologies (UK) announced their latest collaborative effort in ground-breaking brain simulations. They are jointly developing BrainFrame, a novel framework for rapid simulation of large-scale neural models with radically improved programmability.

The ERC Proof-of-Concept project titled “Automated framework for rapid simulations of high-detail brain models” (BrainFrame) funded by the European Commission under Grant Agreement No. 641000 with a budget of 150 thousand Euros reinforces the synergy between experts from neuroscience and high-performance computing who combine their knowledge and methodologies towards next-generation brain-simulation platforms. The goal of BrainFrame is to develop a powerful platform that will allow neuroscientists to easily simulate neural networks of extremely large sizes and high levels of modeling detail using the widely accepted simulator-independent language PyNN. BrainFrame will allow universities and research centers to conduct a plethora of novel computational-neuroscience experiments not possible before due to the excessive programming complexity or the impractically long simulation times involved.

The BrainFrame project consortium aims at brain-simulation machines that are energy-efficient, economically affordable and intuitive to use for a wide range of scientists while at the same time being fully capable of modeling and running biologically meaningful (>1 million) populations of neurons. More information about BrainFrame is available at the project website: www.erasmusbrainproject.com/index.php/themes/brainframe

“Maxeler Dataflow Engines maximise real-time, parallel processing in biologically plausible neuronal networks. This will help us to crack the cerebellar code at a level beyond that of all prior art.”

Prof. Chris de Zeeuw, Department head of Neuroscience Department, Erasmus MC

 

“Key concepts known from our current understanding of human brain operation are exploited by the Maxeler’s Dataflow Engines. This makes it truly exciting to see them bringing key competitive advantage to basic research and the advancement of knowledge, within the top European research area of brain simulation.”

Prof. Georgi Gaydadjiev, VP of Dataflow Software Engineering, Maxeler Technologies

 

About the Neuroscience Department of Erasmus Medical Center

The Neuroscience Department of the Erasmus MC is world-renowned for its contributions in the general-neuroscience field. Since 1998, it holds a staggering record of >200 journal papers with >10% in very high impact journals (Science, Nature group, Neuron). On the technical front, it combines long-standing experience, comprehensive experimental facilities and successfully commercialized lab equipment (see www.neurasmus.com). Moreover, it maintains its technical edge in computational neuroscience and parallel programming through (a) in-house modelers, programmers and engineers, as well as (b) a national network (Medical Delta), and (c) a worldwide network of collaborations and participations (e.g. Human Brain Project, HBP). The department is embedded in the ONWAR, Helmholtz, and MGC research school. For further details please visit also: www.neuro.nl.

About Maxeler Technologies

Maxeler Technologies, Inc. pioneers Maximum Performance Computing. It combines high performance computing (HPC) expertise with custom dataflow computing technology to deliver competitive advantage. Its complete and scalable solutions include hardware platforms, software programming tools and complete applications. Maxeler’s technology enables organizations to speed up processing times by 20-50x, with over 90% reduction in energy usage and over 95% reduction in data center space. For further information please visit www.maxeler.com.