Projects

Computer Architecture and Performance Modeling (CAP)

Performance analysis of computer system architectures exploiting new technology trends using advanced workload characterization and modeling techniques.

Primary Research Category: System Science  
Principal Investigator: Herb Schwetman  



Project Overview

Modeling and Analysis

The CAP group does full-system performance modeling and analysis, using analytical techniques such as queuing models and bounds analysis as well as simulation. The group has built and used detailed, cycle-accurate models as well as high-level abstract models to understand the performance effects of architectural design decisions. The group has worked with systems and applications in many areas of interest to Sun customers, including enterprise software, databases, and scientific computing / High-Performance Computing (HPC) applications.

Workload Characterization

The group has significant experience with application profiling, tracing, and analysis tools. This includes the development of custom tools and methodologies to fit project or workload-specific modeling goals. Tools developed and extended by the group have been adopted by other groups within Sun, and the group contributes improvements to open-source analysis tools.

Recent Projects

The group has performed modeling and characterization as part of the following recent projects:

  • The DARPA UNIC project - the CAP group is designing and modeling the performance of novel Silicon-Photonic interconnects. The group has also contributed to a paper about the project, in the July 2009 Proceedings of the IEEE.

  • The DARPA HPCS project - CAP group worked on architecture performance evaluation and HPC benchmark performance characterization
  • CAP group work on performance analysis of the Sun Java Appserver led to architectural improvements
  • Modeling of the Cloudberry in-memory parallel database project
  • Performed early system-level analysis of benefits of Flash memory in server systems


Technical Reports

  • Productive Petascale Computing: Requirements, Hardware, and Software
    By: Michael Van De Vanter, Alan Wood, Chris Vick, Stuart Faulk, Susan Squires and Lawrence G. Votta Jr.
    Report Number: TR-2009-183
    Jul 9, 2009
  • Quantifying hardware counter sampling error in computer system workload characterization
    By: Lodewijk Bonebakker
    Report Number: TR-2007-163
    Apr 24, 2007

    Latest Sun Patents

  • Conservative shadow cache support in a point-to-point connected multiprocessing node
    (May 1, 2007)
  • System and method for reforming a distributed data system cluster after temporary node failures or restarts
    (Apr 17, 2007)
  • System and method for measuring performance with distributed agents
    (Feb 13, 2007)
  • System and method for dynamic cluster adjustment to node failures in a distributed data system
    (Nov 21, 2006)
  • System and method for coordinating access to data for a distributed application
    (Oct 31, 2006)
    See all 8 patents »


    Team Members

    Herb Schwetman
    Mike McCracken