CUDA training efforts started in Spring 2010 through the graduate
course on High-Performance Computing: Technology, Architecture, and
Algorithms with 6 CUDA compatible old generation graphics cards recycled
from discarded computers in the ECE department. With NVIDIA's support,
UA-CTC will operate with the following objectives:
the undergraduate and graduate curriculum of the Department
of Electrical and Computer Engineering by developing a coordinated
sequence of training materials on GPU computing with CUDA
C/C++ for the computer architecture related courses.
classroom instruction with hands-on, real world design and
troubleshooting experience on multi-core architectures and
software development environments.
strategies to improve the learning curve for the students
and teaching curve for the instructors in their classes on
how to restructure existing algorithms or design new algorithms
in such a way that the program architecture overlaps with
the target GPU architecture.
out to scientific computing community on campus and expose
the potential benefits of the GPU for domain scientists.
Train students to exploit the massive computation power of
the GPUs for solving computationally demanding problems in
many domains such as life sciences, plant sciences, hydrology,
aerospace, and computer science.