High performance Computing

High performance Computing
Services and Availability.

Overview

The Research Computing Services section is tasked with supporting those ICT requirements of researchers across the University which are beyond what is expected to be provided by their faculty. It does this by providing a range of physical resources together with advice, training and consulting in the most effective means of using these resources to further their research needs. Where this results in changed methods or major productivity improvements this may be referred to as "eResearch".

High Performance Computing services

General-Purpose production HPC system
This is our main production system. It caters mostly for users who simply need additional computing power, but has the high-speed interconnect required to support the most demanding parallel applications.
Our current system is a cluster system with 50 dual-Xeon compute nodes. (more info.)

Access:	Any UofM researcher (staff or post-grad) can request an account, and collaborators in projects where the Principal Investigator is from UofM
Availability:	The system is available (via ssh) 24/7, however we reserve the right to schedule downtime for maintenance with at least 7 days notice.

Grid-computing system
Grid computing is an emerging technique to allow groups to share dispersed resources. RCS is collaborating with a number of local groups who are active in Grid development projects. Researchers who are part of other collaborations using Grid computing should contact the section if they wish to be involved in this area.

Access:	Any UofM researcher, active or interested in Grid development, can request an account.
Availability:	The system is usually available (via ssh) 24/7, however, as the Grid is still under development, the system may have extended downtime periods.

Special allocatable system
This was our development system. It is now available for users who need a small dedicated cluster either for testing or because they have special software needs which prevent them using our main system.

Access: Any UofM researcher may apply to use the system.

Availability: As per agreement with the current allocatee.
Test/Development systems
In addition to our cluster systems, RCS has a number of stand-alone systems which are used for testing and development. Researchers interested in access to these systems should contact RCS staff.

Access: Any UofM researcher may apply to use these systems.

Availability: By arrangement.

Data Services

HPC systems disc server
RCS runs a central disc server which is accesable from all our compute services. All our users are given basic access to the home file-system and extra space for large users is available on request.

Access:	Any UofM researcher currently using RCS compute resources.
Availability:	The file-server is available 24/7, however we reserve the right to schedule downtime for maintenance with at least 7 days notice.

DataGrid storage
As part of our Grid development, RCS has two DataGrid storage nodes. Researchers with the requirement to store files which need to be Grid accessable should contact RCS.

Access:	Any UofM researcher may apply to use these systems.
Availability:	The Storage Nodes are available 24/7, however we reserve the right to schedule downtime for maintenance with at least 7 days notice.

Created:	Maintainer: Research Computing Services Information Services The University of Melbourne Disclaimer
Last modified: 13 December 2006 Access: Unrestricted Copyright © 2008 The University of Melbourne.

High performance Computing Services and Availability.

Overview

High Performance Computing services

Data Services

High performance Computing
Services and Availability.