IU Researches will gain access to IU’s new supercomputer this summer and fall. The system, which is the nation’s most powerful university-owned supercomputer and largest disk-based storage facility, will support IU research into new discoveries in informatics, the sciences and the humanities.
The supercomputer features a new distributed shared-memory cluster, called Draco, consisting of 512 IBM JS21 Bladeservers, each with two dual-core PowerPC 970MP processors, connected via gigabit ethernet to a 200 TB GPFS filesystem. Batch jobs will be managed with IBM’s Loadleveler and the Moab Workload Manager. The new system will run Linux, and most users should have no difficulty recompiling codes that they have used on AVIDD. More processing power is also being added to the Libra cluster that will facilitate large memory jobs with access to an additional 50TB volume.
IU Researches will gain access to IU’s new supercomputer this summer and fall. The system, which is the nation’s most powerful university-owned supercomputer and largest disk-based storage facility, will support IU research into new discoveries in informatics, the sciences and the humanities.
The supercomputer features a new distributed shared-memory cluster, called Draco, consisting of 512 IBM JS21 Bladeservers, each with two dual-core PowerPC 970MP processors, connected via gigabit ethernet to a 200 TB GPFS filesystem. Batch jobs will be managed with IBM’s Loadleveler and the Moab Workload Manager. The new system will run Linux, and most users should have no difficulty recompiling codes that they have used on AVIDD. More processing power is also being added to the Libra cluster that will facilitate large memory jobs with access to an additional 50TB volume.
Researchers can also take advantage of an innovative new “data capacitor” system that is being developed to help deal with the massive amounts of data generated over very short time intervals by advanced digital instruments.
More technical detail is available at:
http://racinfo.indiana.edu/spotlight.shtml
Questions or comments may be sent to Research and Academic Computing at
rac@iu.edu.