Sparse Matrix Multiplication on a Field-Programmable Gate Array

To extract data from highly sophisticated sensor networks, algorithms derived from graph theory are often applied to raw sensor data. Embedded digital systems are used to apply these algorithms. A common computation performed in these algorithms is finding the product of two sparsely populated matrices. When processing a sparse matrix, certain optimizations can be made by taking advantage of the large percentage of zero entries. This project proposes an optimized algorithm for performing sparse matrix multiplications in an embedded system and investigates how a parallel architecture constructed of multiple processors on a single Field-Programmable Gate Array (FPGA) can be used to speed up computations