The presentations will be available on CD. If you would like one of these CDs, please contact
Bernadette Watts (

Revised April 14, 2003
List of Acronyms
Acronym Finder

General Guidelines (These instructions were given as guidelines to participants):

  1. We will assume that everyone will read the supplied bio. There would be no time wasted in lengthy introductions. Instead, self-introduction is strongly encouraged in which you can share some personal information to establish a better rapport with the audience. Share a secret, a hobby, a joke, an embarrassing moment, or an incident.
  2. About one-third time is planned for audience participation in each session.
  3. Session Chairs, Moderators, and Panelists will have 10 minutes each for their prepared talks.
  4. Keynote and other speakers will have 20 minutes for talk and 10 minutes for Q&A.
  5. Session chair will decide the order of the presentations.
  6. Each speaker is encouraged to email the presentation ahead; at the latest we will copy it on our flash card at the workshop.
  7. Times allotted will be strictly observed.

Monday, March 3, 2003

4:00pmGuest Check-inFront Lobby
5:30pm – 7:30pmWelcome
Buffet Reception
Animas Room
Tuesday, March 4, 2003
7:00am – 7:00pmRegistration Work StationLobby/Break out area
7:00am – 7:30amContinental BreakfastInnisbrook Foyer
7:30am – 8:00amWelcome & Intro– Bill Camp (SNL), Welcome,
Neil Pundit (SNL), Technical Highlights, and
David Womble (SNL) Welcome and CSRI
8:00am – 8:30amKeynote Speaker: Jeff Nichols (ORNL)
Substituted for Thomas Zacharia.
The Challenge and Promise of Scientific Computing
8:30am – 9:00amKeynote Speaker: Bill Camp (SNL)
Architectural Considerations for Petaflops and Beyond
9:00am – 9:15amBreakInnisbrook Foyer
9:15am – 11:30amGeneral Session
Report on Recent Workshops & Conferences
Chair: Fred Johnson (DOE Office of Science)

Q1: What was the MOTIVATION for the workshop/conf? Burning issues?
Q2: What are 3 significant results or conclusions?
Q3: Was there consensus in the conclusions reached?

ASCI HPC Linux Meeting, Sonoma, CA – Nov, 2002 
Norman Morse (DOE HQ)

Sandia Petaflops Workshop Overview for SOS7
Erik DeBenedictis (SNL)

Barney Maccabe (UNM)

Report on 2002 Fault Tolerance Workshop
Patty Hough (SNL)

Cluster Performance Enhancements High-Intensity Retreat, September 3-5, 2002, Cloudcroft, NM
Rolf Riesen (SNL)

EuroPVM/MPI 2002
Dieter Kranzlmüller (Univ.of Linz)

Report on Communication Architecture for Clusters (CAC) Workshop
D K Panda (OSU)

Clusters, Fault Tolerance and Other Thoughts
Dan Katz (JPL/Caltech)

Center for Programming Models for Scalable Parallel Computing: Project Meeting Report
Bill Gropp (ANL)

Scaling to New Heights
Mike Levine (PSC)
11:30am – 4:00pmTime blocked for side meetings and/or private activities. 
4:00pm – 6:00pmGeneral Session
Machines Already Operational
Chair: Ed Oliver (DOE, Director of ASCR)
Technical Facilitator: Rolf Riesen (SNL)

Summarization of Main Technical Details for Each Machine
Prepared by Rolf Riesen

Q1: Is your machine living up to the performance expectations? If yes, how? If not, what is the root cause?
Q2: What is the MTBI (mean time between interrupts)? What are the topmost reasons for interrupts? What is the average utilization rate?
Q3: What is the primary complaint, if any, from the users?

Operational Machines: ASCI White
Mark Seager

The ASCI Q System at Los Alamos
John Morrison

NCSA Terascale Clusters
Dan Reed

NSF’s Terascale Computing System
Mike Levine

NOAA Forecast System Laboratory Cluster
Leslie Hart

Jean Gonnord
 5:30pm – 5:45pm BreakPre-function Rm
6:00pm – 6:30pmKeynote Speaker: Ralf Gruber (ETH/EPFL)
Is a Grid Cost-effective?
6:30pm – 7:00pmKeynote Speaker: Jack Dongarra (ORNL and U of Tenn)
Self Adapting Numerical Software (SANS-Effort) for Scientific Computing
7:00pm – 7:30pmKeynote Speaker: Dan Reed (NCSA)
The Petascale Mountain Pass
7:30pm – 9:30pmDinner & Keynote SpeakerKiva Room
 8:45pm – 9:15pm
Keynote Speaker: 
Thomas Sterling (JPL/Caltech)
Exaflops: The Final Frontier
Kiva Room

Wednesday, March 5, 2003

7:00am – 7:00pmRegistration Work StationLobby/Break out area
7:00am – 7:30amContinental BreakfastInnisbrook Foyer
7:30am – 9:30amGeneral Session
Planned Machines
Session Chair: Bob Meisner (ASCI Program, DOE HQ)
Technical Facilitator: Rolf Riesen (SNL)

See the summarization of the main technical details for each machine, prepared by Rolf Riesen, under "Machines Already Operational session."

The Panelists will address the following questions for their machines:
Q1: What is unique in structure and function of your machine?
Q2: What characterizes your applications? Examples are: Intensities of message passing, memory utilization, computing, IO, and data.
Q3: What prior experience guided you to this choice?
Q4. Other than your own machine, for your needs what are the best and worst machines? And, why?

Planned Machines: ASCI Purple, ALC and M&IC MCR 
Mark Seager

SNL Red Storm
Jim Tomkins

Pink – A 1024 Node Science Appliance
Ron Minnich

HPCS2 (AKA the HP Cluster at PNNL)
Scott Studham

Cray X1 and Black Widow at ORNL Center for Computational Sciences
Buddy Bland

The IBM BlueGene/L Machine
José Moreira (IBM Watson Center)

Planned Machines: Blue Planet
Brent Gorda
 9:00am – 9:15am BreakInnisbrook Foyer
9:30am – 11:00amGeneral Session
(Other) Killer Applications
Chair: Ken Washington (SNL)

Killer Apps: Neutrinos’ Revenge
Tony Mezzacappa (ORNL)

Supercomputing and the New (Quantitative) Biology 
Grant Heffelfinger (SNL)

Weather Forecasting on Distributed Memory Supercomputers
Leslie Hart (FSL/NOAA)
11:00am – 11:30amKeynote Speaker: Mike Levine (PSC)
The NSF’s Terascale Computing System and Teragrid: Support for Scientific Research
11:30am – 4:00pmTime blocked for side meetings and/or private activities 
4:00pm – 4:30pmKeynote Speaker: Fred Johnson (DOE SC)
Computational Challenges and Directions in the Office of Science
4:30pm – 5:00pmKeynote Speaker: Bill Gropp (ANL)
The Future of MPI
5:00pm – 7:15pmGeneral Session
Runtime Systems, Operating Systems, Message Passing, and Resource Management
Moderator: Ron Brightwell (SNL)Q1: What works?
Q2: What is unsatisfactory?
Q3: What lies ahead?

Some things work…Some things could be better…What lies ahead…
Stephen Scott (ORNL)

Commodity Tools
Pete Wyckoff (OSC)

I/O on Clusters
Rajeev Thakur (ANL)

Job Scheduling
P. Sadayappan (OSU)

Performance Analysis in the Context of Layered Middleware
Thomas Stricker (Swiss Institute of Technology, ETH)

Advances in the Scyld Beowulf System: The Third Generation
Don Becker (Scyld)

The Good, The Bad, The Future
Jim Harrell (USI)
 5:30pm – 5:45pm BreakPre-function Rm
8:00pm – 9:30pmDinner @ Steamworks RestaurantDowntown Durango

Thursday, March 6, 2003

7:00am – 7:00pmRegistration Work StationLobby/Break out area
7:00am – 7:30amContinental BreakfastInnisbrook Foyer
7:30am – 9:45amGeneral Session
Research and Applications in Europe
Chair: Pierre Kuonen (EIA-FR)
Design and Implementation of a Single System Image Operating System for High Performance Computing on Clusters
Christine Morin (INRIA/IRISA)

Optical Trends in Interconnects
Donald Luijten (IBM’s Zurich Research Lab)

The Intercept Project: A Cluster Using the Full PCI-X Performance
Anton Gunzinger/David Mueller

File System Expertise at SCS
Martin Frey (SCS)
 9:00am – 9:15am BreakInnisbrook Foyer
9:45am – 11:30amGeneral Session
Vendors Contribution
Chair: Jean Gonnord (CEA – French Atomic Energy Commission)

Q: What will major vendors do for supercomputing in this decade?

(listed in alphabetical order by company)
What will Cray do for supercomputing in this decade?
Asaph Zemach (Cray)

Dell High Performance Cluster Computing: An Overview 
Jenwei Hsieh (DELL)

Some Thoughts on HP & HPTC & IA-64
Richard Kaufmann (HP)

Intel’s Contribution to HPC This Decade
Stephen Wheat (Intel)

SGI Contributions to Supercomputing by 2010
Steve Reinhardt (SGI)

Sun’s HPC Approach
John Gustafson (SUN)
11:30am – 4:00pmTime blocked for side meetings and/or private activities 
4:00pm – 4:30pmKeynote Speaker: David Nicol (Dartmouth)
Network Security Research Using High Performance Simulation
4:30pm – 5:00pmKeynote Speaker: Rob Leland (SNL)
Technology Confluence and National Security
5:00pm – 5:45pmBreakPre-function Rm
5:45pm – 7:15pmGeneral Session
Crystal Ball
Moderator: Thomas Sterling (JPL/Caltech)Q1: What does the future hold for supercomputing?
Q2: What do we want to see?

Crystal Ball Panel
Al Geist (ORNL)

Can academic ideas and new concepts still make a difference in high performance computing?
Thomas Stricker (ETH)

SOS7 – Crystal Ball or a Random Walk Through Mike’s Brain
Mike Merrill (NSA)

Bill Camp (SNL) – no formal presentation

Crystal Ball Panel: The Futures of Supercomputing
Bill Gropp (ANL)
7:15pm – 7:30pmThank You – Bill Camp, Host (SNL)Innisbrook
7:30pm – 9:30pmBuffet DinnerKiva Room