Project Description

This project will implement a facility to compress traces using domain-specific compression algorithms. The goal is to reduce data storage required by traces by a factor of at least 2 beyond what is possible using traditional compression tools. In addition, the speed of trace reading should equally be improved or at least be equivalent to existing tools. The project will entail a significant programming portion. The outcome of the project should be a study showing how the algorithm works on standard benchmarks in the area of computer architectures.

Prerequisites

1. Excellent programming skills in C++ (preferred) or Python.
2. Solid knowledge of object-oriented programming.
3. Familiarity with traditional computer organization and system architecture.
4. Knowledge and practical use of traditional data structure, including hashing, sorted lists, balanced trees.
5. Familiarity with basic compression algorithms such as Huffman encoding, Ziv-Lempel, bit-run, etc.

Project Deliverables

1. Working compression code.
2. Report of an algorithm.