Solve Ax=b on FPGA

• Main Author: Liu, Huan Yu
• Other Authors: Ling Keck Voon
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/63414
• Institution: Nanyang Technological University

Significant improvements on Field Programmable Gate Arrays (FPGAs) have been developed recent years. Consequently, solving linear equation system, which is one type of the scientific computation, has been accelerated by the development of hardware. In many scientific fields, linear equation systems are commonly used, and the variables of the system can be from a range of a few to millions. However, for the linear system with high dimensions, very long computational time is needed to solve it. In the past years, many methods for the solution of linear system have been proposed [1]. The principal objective of this project is to solve Ax=b on FPGA board, in which Ax=b is a system of linear equations. To solve this linear system of equations (LSE), Gaussian elimination with partial pivoting method was used in the programming algorithm. As an embedded system, FPGA has limited hardware resources. So, to find a most efficient programming design is important, which means the FPGA size could be smaller and the computing speed could be as fast as possible.