COURSE 2010

“High-Performance Scientific Computing on the GPU with emphasis on CAD and FEA”

by

Prof. K. Suresh
University of Wisconsin, Madison, USA

ABSTRACT



The objective of this short-course is to explore the use of graphics programmable units (GPUs) for accelerating scientific computations. Specifically, we shall consider state-of-the-art CUDA-enabled NVidia GPUs to address some of computational challenges associated with geometric modeling and finite element analysis. The short-course consists of the following five lectures (additional lectures will be delivered, if necessary):


• Introduction of GPU: Monday, February 01, 2010 at 3:30 pm


• CUDA programming: Monday, February 08, 2010 at 3:30 pm


• CUBLAS and FFT on the GPU: Monday, February 15, 2010 at 3:30 pm


• Geometric computations using CUDA: Monday, March 01, 2010 at 3:30 pm


• Finite element analysis using CUDA: Monday, March 08, 2010 at 3:30 pm


• Biomechanics applications on the GPU: Monday, March 15, 2010, 3:30 pm


This is a hands-on course where the students are encouraged to ‘play around’ with GPU programming; familiarity with Matlab and C programming is the only pre-requisite.