Contact Information

Peter F. Sweeney
Research Staff Member
Thomas J. Watson Research Center
      +1dash914dash945dash4318


Research Interests

I consider myself an experimental computer scientist who builds things. My current research interests are in performance analysis and tuning of computer systems  in the cloud.  Previously, I have done research in performance issues of native and Java applications running on multicore systes, the static analysis and optimization of object-oriented languages and online feedback-directed optimization.

I am a member of the Dynamic Optimization Group in the Program Technologies Department working on the WAIT project.

Previous projects:

  • MARS: performance optimization of Java enterprise applications, with a focus on IBM's FileNet Content Manager.
  • THOR: visualization and analysis tool for Java application running on multicores (part of the internal IBM multicore SDK tool suite).
  • Continuous Program Optimization: developed an automatic technique to selectively map data structures to large pages (incorporated into IBM AIX operating system).
  • Jikes RVM (aka Jalapeno) Adaptive Optimization System: dynamic compilation techniques for Java (available as part of Jikes RVM at http://jikesrvm.org/).
  • Java Application Extractor (JAX): a static analysis tool that eliminates unused classes, methods and instance variables (released on alphaWorks).
  • High Performance Fortran compiler: available as IBM product on SP2 distributed super computer.

 

Education

I received my M.S. in Computer Science in May, 1988 from Columbia University's School of Engineering and Applied Science. I received my B.S. in Computer Science in January, 1984 from Columbia University's School of General Studies. I am ABD from the PhD program in the Computer Science Department at the University of Wisconsin in Madison.

Professional Activies