Abstract

Compiler infrastructures are often an area of high interest for research. As the necessity for digital information and technology increases, so does the need for an increase in the performance of digital hardware. The main component in most complex digital systems is the central processing unit (CPU). Compilers are responsible for translating code written in a high-level programming language to a sequence of instructions that is then executed by the CPU. Most research in compiler technologies is focused on the design and optimization of the code written by the programmer; however, at some point in this process the code must be converted to instructions specific to the CPU. This paper presents the design of a simplified CPU architecture as well as the less understood side of compilers: the backend, which is responsible for the CPU instruction generation. The CPU design is a 32-bit reduced instruction set computer (RISC) and is written in Verilog. Unlike most embedded-style RISC architectures, which have a compiler port for GCC (The GNU Compiler Collection), this compiler backend was written for the LLVM compiler infrastructure project. Code generated from the LLVM backend is successfully simulated on the custom CPU with Cadence Incisive, and the CPU is synthesized using Synopsys Design Compiler.

Publication Date

8-2017

Document Type

Master's Project

Student Type

Graduate

Degree Name

Electrical Engineering (MS)

Department, Program, or Center

Electrical Engineering (KGCOE)

Advisor

Mark Indovina

Advisor/Committee Member

Sohail A. Dianat

Campus

RIT – Main Campus

Download

Share

COinS