Description
Traditionally, applications run on large-scale grid infrastructures like EGEE have been loosely coupled, i.e. single jobs or sets of jobs with no need to communicate. This model is sometimes referred to as "capacity" or "high-throughput" computing. In contrast, high-performance computing (HPC) applications typically involve much communication between sub-tasks, and benefit greatly from specialised low-latency interconnects. In this session we will focus on bridging the gap between these two models.
Kiril Dichev
(Hih Performance Computing Centre Stuttgart)
22/09/2008, 17:30
Mr
Bernhard Schott
(Platform Computing)
22/09/2008, 18:10
Dr
Peter Kunszt
(CSCS)
22/09/2008, 18:30
Strengths of HPC and grid communities. Challenges to greater collaboration. Opportunities for the future.
