Floating Point Matrix Multiplication on a Reconfigurable Computing System

Matrix multiplication is one of the most fundamental and computationally intense operation that is used in a variety of scientific and engineering applications. There are many implementations of this normally O(n3) operation. These implementations differ mainly in terms of algorithms or the platforms. In this paper we present our experimentation of using a reconfigurable computing platform for calling such a routine. This routine use our own developed IEEE-754 compliant double precision hardware library elements implemented on our own developed FPGA based reconfigurable platform to provide acceleration.

[1]  V. Strassen.  Gaussian elimination is not optimal , 1969 .

[2]  Viktor K. Prasanna,et al.  Reconfigurable computing systems , 2002 .

[3]  Joseph F. Traub.  Complexity of Sequential and Parallel Numerical Algorithms , 1973 .

[4]  Scott Hauck,et al.  Reconfigurable computing: a survey of systems and software , 2002, CSUR.

[5]  Abbes Amira,et al.  An FPGA based parameterizable system for matrix product implementation , 2002, IEEE Workshop on Signal Processing Systems.

[6]  Viktor K. Prasanna,et al.  Scalable and modular algorithms for floating-point matrix multiplication on FPGAs , 2004, 18th International Parallel and Distributed Processing Symposium, 2004. Proceedings..

[7]  Karl S. Hemmert,et al.  Closing the gap: CPU and FPGA trends in sustainable floating-point BLAS performance , 2004, 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines.