Members of HiPEAC granted the first place in the Open Hardware Competition 2017 organized by Xilinx (Pynq category)
Members of HiPEAC granted the first place in the Open Hardware Competition 2017 organized by Xilinx (Pynq category)
ECE NTUA students Elias Koromilas and Ioannis Stamelos, supervised by Research Associate Christoforos Kachris and Prof. Dimitrios Soudris, developed a novel platform for the hardware acceleration of machine learning applications based on the widely used Spark framework.
The Spynq platform allows the acceleration of Spark applications by utilizing seamlessly reconfigurable hardware accelerators in the same way that software packages are used. The novel platform allows Spark developers to instantiate hardware accelerators without major changes in the original code. Spynq platform allows up to 2x faster execution time compared to high performance processors and is up to 20x more energy efficient compared to contemporary server processors. The platform was developed from the Microprocessors Lab (ICCS/NTUA). The Spynq platform was developed under the Horizon 2020 VINEYARD project.
VINEYARD aims to build energy-efficient heterogeneous data centers based on novel programmable hardware accelerators (namely Dataflow engines and FPGA-coupled servers) that can speedup cloud computing and data analytic applications. Also, VINEYARD is developing a high-level programming framework that allow end-users to seamlessly utilize these accelerators in heterogeneous computing systems by employing typical data-center programming frameworks (i.e. Spark).
The platform is available at github: https://github.com/vineyard2020/SPynq
A video showing the main novelties of this platform is shown here:
https://www.youtube.com/watch?v=ix1NI7yq8O4
More info at VINEYARD project: http://vineyard-h2020.eu
