HPC Centres and Strategies for Advancing Computational Science in Academic Institutions

-- RobAllan - 19 Apr 2010

Planes over Europe - here is the current view.

Please note that the event this week is now cancelled. None of the overseas participants, including Dan and Gabrielle can make it and some of the UK participants are still stuck overseas. We will resume plans for the event on 2-3rd June, please see below. Thank you for your cooperation.

Please note there are currently no flights in the UK because of the Iceland volcano plume, see: http://www.metoffice.gov.uk/corporate/pressoffice/2010/volcano.html . We believe all trains and ferries from the rest of the EU are now full.

e-Science Institute URL for June event - includes registration page for invited attendees.

e-Science Institute URL for April event - includes registration page for invited attendees. Initial invitations have been sent.

Description

Computational science, interdisciplinary research and the use of information technology are transforming education, research, and institutional culture in academic institutions. Advancing the role of computation at universities necessitates attention to providing cyberinfrastructure and HPC facilities, support for interdisciplinary research, and developing computational culture and expertise. HPC centres provide one mechanism for building out a computational initiative, and this workshop series will examine how such centres can contribute to this agenda, how they fit into national and international cyberinfrastructures and initiatives, and their challenges and needs for success

This workshop series will bring together HPC centre representatives, academics and policy makers to discuss the role of university based HPC centres in the advancement of computational science research, education and culture at academic institutions and how HPC centres fit into national and international cyberinfrastructures. A resulting report targeted at influencing the future goals and roadmaps of agencies such as EPSRC and JISC will identify findings and recommendations for building, assessing and sustaining such centres. The workshops will also contribute towards establishing a UK initiative that mirrors the NSF Cyberinfrastructure for 21st Century Science in the USA and could provide input to both the US initiative and the European Commission.

Agenda for First Meeting 21-22/4/2010

The workshop will be organized around discussion sessions with three different focuses that will be attended by all participants. In addition, workshop participants will be divided into teams across three cross-cutting themes. Two sessions will be held for each focus, with the first discussing ideas and the second concentrated on providing findings and recommendations. The cross-cutting theme teams will meet twice during the meeting, and at the end of the meeting each theme will present the implications for the theme. The meeting will close with a discussion of plans for the second workshop.

Focuses:

1) Goals and Assessment

  • what should an effort to advance computational science achieve in an academic environment?
  • what should be the goals of an HPC center?
  • how should success or failure be measured?
  • what are reasonable timescales for success?
  • what are the challenges?

2) Models and Operations

  • what is the optimal operational model for an HPC center?
  • what are the different roles of local, regional and national HPC centers?
  • how can HPC centers be sustained, how should they evolve?
  • what level and what type of support personnel are needed?
  • where should the center fit in the overall university structure and administration?

3) Cyberinfrastructure

  • what are the challenges for advancing cyberinfrastructure in a university setting?
  • how should an HPC center integrate with national and international cyberinfrastructure?
  • what balance of compute, network, storage, visualization and collaboration infrastructure is needed?

Themes:

  1. Research: How will research be enabled and integrated with the HPC center? How are the needs of researchers with different computational experience and requirements met? Is innovative and transformational research encouraged? It there a well structured mechanism for research to drive the mission of the HPC center? What is the involvement of faculty, postdocs, and students with the center?
  2. Education: What role will the center have in the education and training of undergraduate and graduate students?
  3. Interdisciplinary and Non-traditional fields: How will the center encourage and support interdisciplinary research? How will the center support research in non-traditional computational fields such as the arts and humanities?

Day 1

1-1.15 Talk from organizers
1.15-1.45 Intro Talk - Gabrielle Allen - experiences at CCT
1.45-3.00 SESSION 1 - Goals and Assessment
3.00-3.30 BREAK
3.30-4.45 SESSION 2 - Models and Operations
4.45-6.00 SESSION 3 - Cyberinfrastructure

DINNER: SEATED BY THEME

Day 2

9.00-10.00 SESSION 1 - Goals and Assessment
10.00-11.00 SESSION 2 - Models and Operations
11.00-11.30 BREAK
11.30-12.30 SESSION 3 - Cyberinfrastructure
12.30-2.00 WORKING LUNCH: THEMES MEET
2.00-2.30 THEME 1 DISCUSSION - Research
2.30-3.00 THEME 2 DISCUSSION - Education
3.00-3.30 THEME 3 DISCUSSION - Interdisciplinary and non-traditional fields
3.30-4.00 PLANS FOR NEXT MEETING

Also associated with the workshop will be the talk, with date and time TBD:

eScience Infrastructure and the Changing Culture of Research: Successes and Lessons Learned in Australia

Bill Applebe, CEO, VPAC

Traditional research was driven by individual researchers and teams, their insights and inventions, and academic institutions were set up to foster this culture. Research infrastructure, such as instrumentation or computing, was acquired on a competitive grant basis and owned by the researchers. However, some research disciplines, such as Astronomy or High-Energy Physics, have for many years had an "infrastructure-driven" approach to research - what collaborative research can be done with access to the available infrastructure (such as the Hubbell Space Telescope or LHC). This infrastructure and collaboration based approach to research is now rapidly permeating other disciplines such as Bioinformatics, Climate, and even Economics. It is a paradigm shift that is changing the way that research is funded, conducted, and rewarded.

Australia, and the state of Victoria in particular, have been quite successful over the past decade in eResearch collaboration and outcomes, thus attracting major industry, state, and national funding. State investment in eResearch infrastructure has been over $200M in the past few years, matched by national and University investments. This talk will discuss recent Australian successes, lessons learned, organizational structures that facilitate eResearch, and the evolving landscape.

Organizing Committee

Daniel S. Katz University of Chicago, USA dsk@ci.uchicago.edu
Robert J. Allan Daresbury Laboratory, UK robert.allan@stfc.ac.uk
Gabrielle Allen Louisiana State University, USA gallen@cct.lsu.edu

Background

Statement from NSF Cyberinfrastructure for 21st Century Science and Engineering. Office of Cyber Infrastructure

Programme

  1. First meeting at e-SI 21-22/4/2010 followed by writing on 23/4/2010, see http://www.nesc.ac.uk/esi/events/1069/
  2. Feedback to HPC-SIG at next meeting, Cardiff, 28/4/2010
  3. NGS at their meetings or through contacts
  4. NSF Workshop on High Performance Computing Center Sustainability - May 3-5, 2010 (Dan/Gab will provide info on first UK workshop to this US group)
  5. There are also other meetings scheduled in USA
  6. Second meeting at e-SI 2-3/6/2010 followed by writing on 4/6/2010
  7. Feedback as above
  8. Preparation of final report

Report

to be written...

Topic revision: r2 - 30 Apr 2010 - 13:44:01 - RobAllan
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback