Leading University Joins Prominent Network of Institutions Focused on Parallel Computing

The university was selected for its pioneering use of GPU computing and the CUDA programming model across research and teaching efforts within multiple science and engineering departments.

CUDA is NVIDIA’s computing architecture that enables its GPUs to be programmed using industry standard programming languages and APIs, opening up their massive parallel processing power to a broad range of applications beyond graphics.

Other CUDA Centers of Excellence in the U.S. and abroad include Cambridge University, Chinese Academy of Sciences, Harvard University, National Taiwan University, Tokyo Institute of Technology, Tsinghua University, University of Illinois at Urbana-Champaign, University of Tennessee and University of Utah. More than 300 universities worldwide teach the CUDA™ programming model within their curriculum.

“Maryland was one of the first universities to start integrating the use of GPUs and the CUDA architecture into our courses and research,” said Amitabh Varshney, Professor of Computer Science at University of Maryland. “The CUDA programming model is an extremely effective educational tool for students learning parallel programming and no other technology available today provides as powerful and affordable platform for our research as the GPU.”

Researchers at the University of Maryland have been exploring the use of GPUs for general-purpose computing for the past five years, when they have demonstrated how to map a number of problems in science, engineering, and medicine to GPUs. Maryland researchers have also published papers that use the CUDA architecture of NVIDIA GPUs to enable entirely new computational techniques in these disparate fields, ranging from the astrophysical simulation of colliding black holes to the real-time analysis of the acoustic properties of concert halls.

The CUDA Center of Excellence at University of Maryland will support several new projects that make extensive use of GPUs such as DNA sequencing. There has been a dramatic increase in the volume of sequence data that can be analyzed, thanks to GPUs, and sequence alignment programs such as MUMmer, a system developed by University of Maryland with the support of the National Institute of Health, have proven essential to this process. By structuring the required processing in parallel on a GPU, MUMmerGPU achieves more than a 10-fold speedup over a serial CPU version of the sequence alignment kernel. MUMmer GPU is available today through NVIDIA’s Tesla Bio Workbench initiative.

Visit the CUDA Center of Excellence program pages for more information.