GPU Parallel Programming using CUDA

On 19 and 20 of March 2009, a two day Crash Course on GPU Parallel Programming using CUDA was successfully organized by professor St. Kaxira and the Computer Architecture Group of the ECE dept of the University of Patras in cooperation with the PDS Group. Main invited speaker of the seminar was Richard Membarth from the University of Erlangen-N├╝rnberg, Germany. The seminar included in depth presentation of the environment and the programming model of NVIDIA CUDA and detailed discussion concerning the various ways that CUDA can be used in the context of high performance parallel processing. On the second day, a hands-on workshop took place at the seminar room of the Computer Center at the CEID department. Cutting edge hardware platforms were utilized, including 2 NVIDIA Tesla S1070 1U units by the PDS Group and the HPCLAB (CEID), 6 NVIDIA GTX295 nodes by the CAG group (ECE), and 3 NVIDIA GTX280 nodes by the CSL lab (ECE).


March 19 - Day 1
Seminar on GPU Parallel Programming using CUDA
University of Patras Conference Center: 10:00 - 13:00 March 20 - Day 2
Hands-on Workshop on GPU Parallel Programming using CUDA and NVIDIA high-end GPUs.
Seminar room - Computer Center of the Computer Engineering & Infromatics dept: 10:00 - 18:00