Speaker
Dr
Maxim Potekhin
(BROOKHAVEN NATIONAL LABORATORY)
Description
The simulation program for the STAR experiment at Relativistic Heavy Ion Collider at
Brookhaven National Laboratory is growing in scope and responsiveness to the needs of
the research conducted by the Physics
Working Groups. In addition, there is a significant ongoing R&D activity aimed at
future upgrades of the STAR detector, which also requires extensive simulations
support. The principal computing facility used by STAR to conduct the simulations
studies is a farm containing 400 nodes, with a total of 1000 CPUs.
OpenScience Grid (OSG) resources have been successfully used in the past and
routinely used in STAR. However, the explosive growth of the computing power and the
rapid evolution of the distributed computing landscape demand for the STAR
Collaboration to dictate that all available options are considered, from Open source
to commercial grids using a thin modular layer interfacing with the many “grids”. Sun
Grid from Sun Microsystems aims to deliver enterprise computing power and resources
over the Internet, enabling developers, researchers, scientists and businesses to
optimize performance, speed time to results, and accelerate innovation without
investment in IT infrastructure.
We have successfully run a part of our production jobs on the SunGrid facility and
will present our experience with its interface, performance and related issues and
discuss ongoing efforts and development to interface it with the STAR Unified
Meta-schedule (or SUMS).
| Submitted on behalf of Collaboration (ex, BaBar, ATLAS) | STAR |
|---|
Primary authors
Dr
Jerome LAURET
(BROOKHAVEN NATIONAL LABORATORY)
Dr
Maxim Potekhin
(BROOKHAVEN NATIONAL LABORATORY)
Co-authors
Dr
Ari SHAMASH
(Sun MicroSystem)
Mr
Gabriele CARCASSI
(Sun MicroSystem)
