Changhao Jiang is a fourth year Ph.D. student at the Computer Science Department of University of Illinois at Urbana Champaign. Prior to his Ph.D. study, he had worked for two years at the Computer Science Department of Carnegie Mellon University. He received both his Bachelor and Master degrees in Computer Science from Tsinghua University, China in 1999 and 2001 respectively. He is also nicked as "Joe" by colleagues and friends.

• Automatic Tuning Matrix Multiplication on Graphics Hardware, Changhao Jiang and Marc Snir, accepted by the Fourteenth International Conference on Parallel Architecture and Compilation Techniques (PACT), Saint Louis, Sep, 2005
• Programming Patterns for Architecture-Level Software Optimizations on Frequent Pattern Mining, Mingliang Wei and Changhao Jiang and Marc Snir, accepted by IEEE 23rd International Conference on Data Engineering (ICDE), Istanbul, Turkey, 2007
• Old publications

• Automatic Tuning Matrix Multiplication on Graphics Hardware, Invited by professor Pat Hanrahan to give a talk at the Computer Graphics Laboratory of Stanford University, Jun, 2005
• General Purpose Computation and Performance Tuning on GPUs, Invited by professor María Jesús Garzarán to give a guest lecture for UIUC CS Course CS498DP "Program Optimization", Oct, 2006

• Ray Ozzie Fellowship (2003-2004)
• Illiac Fellowship (2003-2004)
• Illiac Fellowship (2004-2005)
• University of Illinois Fellowship (2005-2006)

       One way of attacking this problem is to have an automatic software generation system to generate multiple versions of a program and empirically evaluate their performance to pick the best version. A promising research field -- automatic library generation via empirical search, is emerging to follow the route.

       Changhao is working with professor Marc Snir and professor María Jesús Garzarán on automatic library generation via intelligent empirical search. Along this line, he has done some work on automatically tuning matrix multiplication for graphics hardware(GPU), which is the first automatic library generation system on GPUs. Together with his colleagues, he conducted a systematic study on programming patterns of architecture-level software optimizations for frequent pattern mining, a fundamental problem in data mining. To tackle problems whose performance is not only affected by the underlying architecture but also dependent on the input characteristics, he has made significant progress in the research of using machine learning techniques to select the best algorithm based on input characteristics for data mining tasks, which is the first research attempt of its kind for data mining tasks. The result has been submitted for publication.

       Currently, Changhao is looking into ways of efficiently utilizing new architectures such as the IBM cell processor and more recent GPUs to perform important computing taks.