Spring 2009 Final Project and Resources

The objective of ECE498AL final projects is to provide students with first-hand experience in accelerating real-world applications. These projects should allow students to apply their knowledge in writing high-performance massively parallel applications to applications whose acceleration can enable significant advancement of a research area. The areas of interest include science, engineering, medicine, and financing. In order to save your time in search for a project, we have created a recommended list of projects that are of interest to world-class researchers. These researchers will be interested in help mentoring the student teams pursuing their projects.

Due: 4/09/2009: Please submit your team's Final Project Design Documentation (doc) by April 9th.
Due: 5/06/2009: Final Project Presentations (template slide [ppt].)
Due: 5/08/2009: Final Project Report (template [doc].)

Leaders Project
Dae Hyun Kwon and Richard Tseng Substrate Coupling Analaysis in Integrated Circuits (pdf)
Georgios Tsekenis Dislocation Dynamics Simulation (pdf)
Bob Brunner Cosmology (doc)
Bei Peng and Seungchul Lee Adaptive Polynomial Transversal Filtering (pdf)
Lijuan Luo EDA Processes (doc)
Justing Lee Parallel Ray Tracing (doc)
Rahul Yargop Multivariate Gaussian Distributions for Automatic Speech Recognition (doc)
Emmanuel Amaro and Tomas Zegard Topology Optimization (pdf)
Ross Favero, Tyler Wickenhauser and Simion Venshtain Delauny Triangulation (doc)
Joshua Blackburn and Quang Nguyen Cubic Spline Interpolation (doc)
Resources:
NVIDIA Official CUDA Page http://developer.nvidia.com/object/cuda.html
CUDA Occupancy Calculator A programmer's tool that allows you to compute the multiprocessor occupancy of a GPU by a given CUDA kernel. The multiprocessor occupancy is the ratio of active warps to the maximum number of warps supported on a multiprocessor of the GPU, and is helpful in determining how efficient the kernel will be on the GPU.
MCUDA translator A tool developed by John Stratton that translates CUDA applications into multithreaded C code. We will have the tool available on the NCSA cluster in the coming weeks.
CUDA-tune automatic optimization system A tool developed by the IMPACT group that eases the tediousness of manually generating a testing a large number of kernel configurations. Hopefully to be made available on the NCSA cluster as well. More details as they come.

 

  Archived final project listings from previous semester(s):  
  Spring 2007 - First-time course offering by Prof. Hwu (UIUC) and Prof. Kirk (NVIDIA)!