The purpose of this thesis is to study the methodology of behavioral synthesis and evaluate its usefulness compared to Register Transfer Level (RTL) synthesis. Custom IC design uses high-powered synthesis tools. Engineers have traditionally used RTL level descriptions of their circuits as input to these synthesis tools. As new Behavioral Synthesis tools are becoming more powerful, the option to describe their circuitry in a higher and more abstract level is becoming a more feasible option. Describing circuitry at a higher level has many advantages. It is easier to make architecture changes and higher level descriptions generally have significantly less lines of code and faster development times. To study behavioral synthesis a tri-linear interpolation algorithm is used. An RTL style and two different behavioral styles are used. Each are compared for area, power consumption, synthesis time, code length and throughput. The design is simulated before and after synthesis to verify the accuracy of the design using VHDL. Behavioral Compiler from Synopsys will be used to synthesize the design from VHDL to the gate level. It was found that behavioral synthesis can produce results nearly as good as an RTL described circuit. The results were generally 20% - 30% worse for this implementation using behavioral synthesis.

Library of Congress Subject Headings

Digital integrated circuits--Design and construction; Data structures (Computer systems); Computer architecture; VHDL (Computer hardware description language)

Publication Date


Document Type


Department, Program, or Center

Computer Engineering (KGCOE)


Dianat, Soheil

Advisor/Committee Member

Parrett, Gary


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TK7874.65 .G6 1999


RIT – Main Campus