Utilizing the graphical processing unit for fuzzy logic inference [abstract]

Abstract

The graphical processing unit (GPU) contained on modern video cards can be a powerful tool for performing certain nongraphical computational tasks. This is because the specialized hardware necessary for speeding up graphical processing can also utilized to significantly reduce the time necessary for calculations which can be formated in a similar way to graphical calculations. Specifically,vector/matrix operations and floating point operations are much faster. One area in which this technique can be applied is that of fuzzy logic. The goal of this project is to use the parallel processing and other specialized hardware features of the GPU to perform fuzzy logic inference more quickly than would be possible using a conventional computational methods on the CPU. Both construction of the rule matrices and use of the inference engine will take place on the GPU. The increase in speed resulting from the use of the GPU will depend on the specific hardware on which the system is run, as well as the problem size. As the amount of data increases, the benefits will become more apparent. On problems of sufficient size, the GPU system should perform the calculations at least one order of magnitude.