VINEYARD aims to develop novel frameworks that will allow the efficient utilization of heterogeneous platforms from high-level programming frameworks and simulation platforms. One of the main frameworks that VINEYARD has developed is the BrainFrame. BrainFrame aims to allow the seamless deployment of several hardware accelerators (such as Maxeler’s Dataflow Engines, FPGAs, and Xeon Phi) for highly efficient brain simulation.
The general goal of the project is to apply high performance, innovative solutions enabling large scale, accurate or real-time brain simulations and enhance experimental setups or data analysis of the experimental data concerning brain research. Thus, our activities employ a multitude of HPC and other technologies such as FPGAs, Dataflow Computing, GPUs and Many-core processors. The long-term goal and main focus of the effort within this theme is the development of a generic tooling framework for accelerated brain simulations, the BrainFrame Framework.
More information HERE