Electrical and Computer Engineering ETDs

Publication Date

Fall 12-1-2016

Abstract

When developing a hardware or software system, the problem at hand may lend itself to multiple solutions. During the implementation process for such systems, it can be helpful to prototype multiple versions that use distinct paradigms, and determine the efficiency of each according to some metric, such as execution time. This paper presents a portable, lightweight build system designed for easy benchmarking and verification of competing implementations of an algorithm. Also presented is a sample project that uses this system to compare the performance and correctness of CPU, GPU, and FPGA implementations of a signal recovery algorithm

Keywords

gpu fpga c

Document Type

Thesis

Language

English

Degree Name

Computer Engineering

Level of Degree

Masters

Department Name

Electrical and Computer Engineering

First Committee Member (Chair)

L. Howard Pollard

Second Committee Member

Daryl Lee

Third Committee Member

James Plusquellic

Download

Share

COinS