Date of Award
Master of Science
Current Graphics Processing Units (GPUs) provide programmable vertex and fragment processing circuitry, which has attracted considerable attention from researchers seeking to exploit the power of GPUs for general purpose computing. However it should not be forgotten that even the nonprogrammable hardware of a graphics processor has potential to speed up algorithms while maintaining compatibility with many older and lower end GPUs. Here we utilize standard OpenGL functionality, which has been present for many years on most GPUs, to generate dotplots via a non-traditional method. Dotplots, a graph commonly used in genetic data comparisons to identify sequence structure and similarities between sequences, has a computation time that increases rapidly as the length of the input sequences grows. A GPU can be utilized to achieve significant speedup over a purely CPU based algorithm. As sequence size grows so large that the dotplots are forced to be offloaded to a persistent storage device, we find that this advantage is diminished due to a bottleneck arising from the necessary manipulation and transfer of the generated data. We also consider and discuss coping strategies and consider future possibilities to surmount this problem.
This thesis is only available for download to the SIUC community. Others should
contact the interlibrary loan department of your local library.