Architectural design tool for low area band matrix LU factorization

An area efficient band matrix design tool is presented for computation of LU factorization. Band matrices are used in many applications such as digital signal processing, power system analysis, finite element systems, and many more. In many of these applications, LU factorizations are needed to solve linear equations and calculation of a matrix determinant. Due to the large amount of data that needs to be computed and stored for this factorization, the goal of the proposed design tool is area minimization without compromising speed. Based on the number of zero elements and bandwidth, the design tool uses a High Level Synthesis to create the RTL code and testbench for these factorizations for the desired matrix. The proposed design tool reduces area, required memory location and chip to market time by generating a testbench and doing error analysis during RTL code generation. The generated RTL code is universal and can be used directly for any FPGA and VLSI platforms.

[1]  A. Fujii,et al.  Parallel matrix distribution library for sparse matrix solvers , 2005, Eighth International Conference on High-Performance Computing in Asia-Pacific Region (HPCASIA'05).

[2]  J. Saniie,et al.  HW/SW co-design platform for image and video processing applications on Virtex-5 FPGA using PICO , 2010, 2010 IEEE International Conference on Electro/Information Technology.

[3]  John G. Lewis Implementation of the Gibbs-Poole-Stockmeyer and Gibbs-King Algorithms , 1982, TOMS.

[4]  H. D. Simon,et al.  A spectral algorithm for envelope reduction of sparse matrices , 1993, Supercomputing '93. Proceedings.

[5]  I. Duff,et al.  Direct Methods for Sparse Matrices , 1987 .

[6]  Erdal Oruklu,et al.  Realization of area efficient QR factorization using unified division, square root, and inverse square root hardware , 2009, 2009 IEEE International Conference on Electro/Information Technology.

[7]  Yasuaki Inoue,et al.  High-performance systolic arrays for band matrix multiplication , 2005, 2005 IEEE International Symposium on Circuits and Systems.

[8]  Richard Rosen Matrix bandwidth minimization , 1968, ACM National Conference.

[9]  E. Cuthill,et al.  Reducing the bandwidth of sparse symmetric matrices , 1969, ACM '69.