Sparse Matrix Multiplication on Maxeler Dataflow Engine

Matrices are powerful mathematical data representation tools in the engineering and science fields.Also, Sparse matrices which contains few non zero elements are also widely encountered. Representation of sparse matrices and multiplication of sparse matrices with a vector in an efficient way is also a key issue.
In order to calculate fast and gain speed in that of kind of matrix multiplications, a lot of algorithms and different architectures are developed and used.We try to use a new technology based on FPGAs developed by Maxeler Company. FPGA platforms are widely used for acceleration of computations the Maxeler company produces FPGAs with new concept, as well as high level programming languages and platforms for designing of FPGA architectures.
We try to calculate the sparse matrix multiplication with a vector by using the algorithm developed by Blelloch, G. E., Heroux, M. A., & Zagha, M. (1993) "Segmented Operations for Sparse Matrix" and try on the maxeler data flow engine in order to use huge data sets from the Matrix Market , gain speed and analyze the results.