Doctoral student Johannes de Fine Licht and Bachelor student Manuel Burger from the Scalable Parallel Computer Laboratory have won in two categories at the Xilinx Open Hardware Design Contest.

Two of the six prizes in the Xilinx Open Hardware Design Contest 2020 went to students from the Scalable Parallel Computer Laboratory (SPCL) led by Professor Torsten Hoefler. Doctoral student Johannes de Fine Licht won in the Compute Acceleration category, while undergraduate student Manuel Burger won in the PhD category – for a project he worked on for his Bachelor’s thesis. A third project from Professor Hoefler’s group, by undergraduate student Simon Meierhans, placed among the thirty-two contest finalists.

Johannes de Fine Licht won in the Compute Acceleration category with his submission “Flexible Communication Avoiding Matrix Multiplication on FPGA with HLS”. He has worked on this project since the summer of 2019 and has recently published a external pagepaper on it with co-author Grzegorz Kwasniewski. “Matrix multiplications are ubiquitous in science,” de Fine Licht explains. “But the truly exciting thing about this project and about the competition is its openness. A strong open source movement exists in the software community, but hardware is usually much more closed.”

The competition encourages participants to openly publish their work, which straddles the boundary between software and hardware. Participants write code for reprogramming reconfigurable circuits called field-programmable gate arrays or FPGAs. Making the code freely available allows other scientists to build on it or to use it in their work. “At SPCL, we strive for more openness in the hardware space”, says Johannes de Fine Licht. “This is why I joined the competition and recommended it to the students I was supervising at the time, Manuel Burger and Simon Meierhans.”

By playing the video you accept the privacy policy of YouTube.Learn more OK
Johannes de Fine Licht's project "Flexible Communication Avoiding Matrix Multiplication on FPGA with HLS" won in the Compute Acceleration category.

Manuel Burger’s winning project, “Portable Linear Algebra on FPGA using Data-Centric Parallel Programming”, was originally created for his Bachelor’s thesis. “Participating in the competition raised my motivation another notch”, he says. “Of course I’m proud to have won, especially in the PhD category, but I have also learned a lot and contributed something useful, since my work has applications in high-performance computing, in particular for simulations.” Burger’s project was part of the Data-Centric Parallel Programming (DAPP) effort, which is funded by the European Research Council (ERC) under the European Union’s Horizon 2020 programme. Meanwhile, he has successfully earned his Bachelor’s degree and is beginning his Master’s in Data Science.

By playing the video you accept the privacy policy of YouTube.Learn more OK
Manuel Burger's winning project "Portable Linear Algebra on FPGA using Data-Centric Parallel Programming" was part of his Bachelor thesis, supervised by Johannes de Fine Licht.

The Xilinx Open Hardware Design Contest is a competition in which doctoral and undergraduate students across Europe can showcase their technical and creative skills. Projects are judged on five categories: technical complexity, implementation, marketing/innovation, documentation and reusability. It has been held annually since 2015.

external pageWebsite

JavaScript has been disabled in your browser