Description
Evolvement of US and European HPC Infrastructures (90 mins)
Hermann Lederer, DEISA, Thomas Eickermann, PRACE, and Phil Andrews, TeraGrid
This Session, will start with an actual overview of the DEISA, PRACE, and TeraGrid initiatives, including joint technologies and standards used, their evolving collaboration described, and a joint vision about the future HPC/Grid/Cloud infrastructures for e-Science exhibited.
Over the last 30 years, a few tens of powerful HPC centres have been built and operated in Europe and the US and successfully helped to advance computational sciences and the whole science community. During the last five years, two EU FP6/FP7 funded projects have been established to unite some of the most powerful supercomputing centres into one European HPC ecosystem. Recently, cooperations started between the European initiatives with the US TeraGrid.
The DEISA Consortium has deployed and operated the Distributed European Infrastructure for Supercomputing Applications. In DEISA-2, the consortium continues to support and further develop the distributed high performance computing infrastructure and its services. Activities, services, and standards relevant for Applications Enabling, Operation, and Technologies are continued and further enhanced. Grand challenge projects are performed on a regular basis within the DEISA Extreme Computing Initiative (DECI). By selecting the most appropriate supercomputer architectures for each project, DEISA is opening up the currently most powerful HPC architectures available in Europe for the most challenging projects.
The Partnership for Advanced Computing in Europe (PRACE) prepares the creation of a persistent pan-European HPC service, providing European researchers with access to capability computers and forming the top level of the European HPC ecosystem. The service will comprise three to five superior tier-0 HPC centres strengthened by regional and national supercomputing centres working in tight collaboration through grid technologies. PRACE is preparing for the implementation of this infrastructure in 2009/2010 by defining and setting up the legal and organisational structure involving HPC centres, national funding agencies, and scientific user communities. It is procuring a set of prototypes to prepare for the deployment of the first Petaflop/s production systems including selection, evaluation, benchmarking and scaling of middleware, libraries and applications.
The TeraGrid is the amalgamation of the major U.S. National Science Foundation HPC centers, currently with approximately a dozen Resource Provider sites and a total computational capacity of over 1 Petaflop. This is scheduled to increase to about 2 PF in 2009 and 3 PF in 2010. The TeraGrid itself will experience a major transformation in 2010, with
a new organizational model expected to replace the current approach. This reorganization is presently in the proposal stage. We will discuss the history of the TeraGrid, the past changes in the funding and strategic plans, and any available information on future directions, including technology, standards, and international collaboration.
In this session, special attention will be given to the importance and use of standards for building and operating complex HPC e-Infrastructures, and their important role in disseminating and further developing these standards will be described.
Agenda:
- Introduction, overview, and vision for
HPC ecosystems for e-science at continental
scale
- The DEISA project, presence and future
- The PRACE project, presence and future
- The TeraGrid project, presence and future
- Moderated Q&A and discussions about
success and challenges of standards,
international collaboration, and next
steps.
