Technology of Code Compression for Embedded System

This paper expatiates the redundancy of instruction in embedded system, and analyzes some methods for code compression at present. It also discusses the difference between thumb instruction and other methods of code compression, and explains the essence of thumb. And finally, it proposes a feasible method for code compression which fits fast decoding by table-based design through analyzing the theory and thoughtway of Arithmetic coding. This method not only can be used for any instruction set with fixed-size instructions but also may not break the structure of processor pipeline. And it keeps compression ratio around 70% with its fast execution of decompression.