Agenda

Training Day

Monday, Oct 3, 2005

Oakland Scietific Facility, 415 Thomas L Berkley Way (20th St.) Oakland, CA

Jacquard - NERSC's New 640-CPU Opteron Cluster

Time (Pacific)

Topic

Presenter

9:00

Jacquard Overview: a high-level description of system, processors, interconnect, file systems, compilers, benchmarking, etc.

Richard Gerber

9:45

Jacquard Nodes and CPUs: Opteron basics, differences from POWER 3, node configuration, memory layout, processor affinity

David Skinner

10:30

BREAK

10:45

High Speed Interconnect and MVAPICH: InfiniBand characteristics (latency, bandwidth), network topology, differences from Seaborg, MVAPICH overview

Bill Saphir

11:30

Compiling: Pathscale Fortran, C, C++; mpif90, mpicc, mpicxx, recommended compiler options, useful compiler options, libraries, porting from Seaborg, porting from other Linux clusters

Michael Stewart

12:15

LUNCH in downtown Oakland

2:00

Running Jobs: Batch system, PBS overview, queues, submitting jobs, monitoring jobs, interactive jobs, differences from LoadLeveler

David Turner

2:30

Software overview: user environment, libraries, debuggers and programming tools

Jonathan Carter

3:00

BREAK

3:30

Hands-on: bring your laptop with SSH and X-Windows

Richard Gerber & NERSC Staff

5:00

Adjourn. Machine room tour.

NERSC Staff

NUG Business Meeting

Tuesday, Oct. 4

Berkeley Lab, Building 50 Auditorium

Time (Pacific)

Topic

Presenter

8:00

Welcome & logistics

David Dean

8:15

DOE Update

David Goodwin

8:30

The Greenbook

Stephane Ethier

9:00

Science Driven Computing: NERSC's Five-Year Plan for 2005 - 2010

Horst Simon and Bill Kramer

9:45

Summary of Five-Year Plan Reviewer Comments

Bill Kramer

9:55

Welcome from LBNL Director Steve Chu

10:00

BREAK

10:15

NUG Feedback on 5-year Plan

NUG Members, chaired by David Dean

10:45

NERSC/LBNL Data Analytics Project

Wes Bethel

11:15

Facility Wide File System

Jim Craw

11:30

Proposed OMB Metrics for NERSC

Bill Kramer

11:45

BREAK

12:00 (Lunch Talk)

Progress in Supercomputing: The Top Three Breakthroughs of the Last 20 Years and the Top Three Challenges for the Next 20 Years

Horst Simon

1:15

NERSC INCITE and SciDAC Program Updates

Richard Gerber

1:30

Scientific Computing Highlight I: Simulating the Convective Epoch of Core Collapse Supernovae

Doug Swesty

2:15

Visualization Highlight: Using AVS/Express to model cell membranes from electron tomography data

Cristina Siegerist

2:30

BREAK

2:45

Scientific Computing Highlight II: High-Fidelity Terascale Simulations of Turbulent Combustion

Jackie Chen

3:30

Scientific Computing Highlight III: Feeding the Pipeline: The SNfactory Search for Nearby Supernovae

Richard Scalzo

4:15

Q&A, open discussions

NUG Members, chaired by David Dean

6:00

Working NUG Executive Committee dinner: Future systems and requirements for the next large system.

Jonathan Carter