Date Added: Jun 2010
This paper presents a rigorous methodology for accurately benchmarking the capabilities of an FPGA architecture. The goal of benchmarking is to compare the capabilities of one FPGA architecture versus another. Since the FPGA industry does not conform to a standard benchmarking methodology, this paper describes in detail the methodology employed by Altera and shows how a poor methodology can skew results and lead to false conclusions. The complexity of today's designs together with the wealth of FPGA and computer-aided design (CAD) tool features available make benchmarking a difficult and expensive task. To obtain meaningful benchmarking results, a detailed understanding of the designs used in comparisons as well as an intimate knowledge of FPGA device features and CAD tools is required.