Paradyn Logo

Paradyn Week

Paradyn Logo
March 17-18, 1998
University of Wisconsin
Madison, WI

PostScript copies of presentation materials available by clicking on titles.

Tuesday, March 17, 1998
Technical Session Agenda

08:15 - 08:45 Welcome, Introductions, and Overview
Bart Miller, U. Wisconsin
08:45 - 09:15 Performance of Runtime Instrumentation and its Interaction with Processor Features
Bryan Buck, U. Maryland
09:15 - 09:45 A New Scheme for Performance Consultant Searches in the Code Hierarchy
Matthew Cheyney, U. Wisconsin
09:45 - 10:30 Experiment Management Support for Performance Tuning
Karen Karavanic, U. Wisconsin
10:30 - 10:45 Break
10:45 - 11:30 KernInst: Dynamic Instrumentation of the Operating System Kernel
Ariel Tamches, U. Wisconsin
11:30 - 12:00 Update: Paradyn v2.1 Release PDF
Brian Wylie, U. Wisconsin

12:00 - 13:30 Lunch

Tuesday Afternoon
13:30 - 14:00 Paradyn and Globus
Ian Foster, Noam Freedman & Brian Toonen, Argonne National Lab.
14:00 - 14:45 Instrumenting Threaded Programs
Oscar Naím, U. Wisconsin
14:45 - 15:30 Performance Measurement of Interpreted (Java) Programs
Tia Newhall, U. Wisconsin
15:30 - 15:45 Break
15:45 - 16:15 Runtime Tools Standard Part 1: DynInst API
Jeff Hollingsworth, U. Maryland
16:15 - 16:45 Runtime Tools Standard Part 2: DAIS
Doug Pase, IBM
16:45 - 17:15 Concluding Remarks and Feedback
Jeff and Bart

Tuesday evening, March 17, 1998

18:00 - Dinner

Wednesday, March 18, 1998
DynInst/DAIS Standards Meeting Agenda
1) Process Issues

    A. Is there sufficient interest for the results of this standards effort?
       Or should Maryland, Wisconsin, and IBM proceed as we have?
    B. Who are the participants (active and observing)?
       Do other vendors wish to participate?
       Do other research projects wish to participate?
       What is the level of commitment of resources?
    C. How are decisions made?
2) Scope

    A. What are we trying to standardize?
       A multi-layered model, with well-defined interfaces at the
       dyninst and DAIS levels:

       Areas of functionality:
        * Platform independent process instrumentation on a single node.
        * Platform independent process control functions on a single node.
        * Platform independent extracting data from processes.
        * Multi-node, multi-tool support/RPC architecture.
    B. Who are the target users?
    C. What is the final product?
        * Standard specification document?
        * Reference implementation?
        * Test suite?
3) Time Table

    A. What do we want to have done and when?
4) Technical Discussion of dyninstAPI

    A. Status of current design
    B. Status of initial implementation
    C. Vision for the future
5) Technical Discussion of DAIS

    A. Status of current design
    B. Status of initial implementation
    C. Vision for the future
6) Logistics of current dyninstAPI implementation

    A. Process to incorporate changes from partners?
    B. Who will implement new features as interface evolves?
    C. How to minimize rapid changes in the code base
    D. Testing: who and how?
    E. Legal: ownership and rights to code.
7) Where do we go from here?

    A. Time/place for the next meeting?
    B. Tasks to be done by the next meeting
[Notes from DynInst/DAIS meeting]

Last modified: Sun Mar 15 14:54:25 CST 1998 by bart