Introduction

As a consequence of the application driven nature of development within MeSsAGE Lab we manage and provide access to a small set of service and compute infrastructure including the ARCS Nimrod Portal and the East cluster of the EnterpriseGrid. Our experience within e-Research and Science, and contacts both inside Monash (MeRC , ITS , FIT ) and externally (VPAC , ARCS , PRAGMA , etc) can be particularly helpful to the applications scientist new to grid computing and/or e-Science generally. We can for example often expedite the process of gaining access to shared resources simply by pointing to the appropriate contact.

East Enterprise Grid

Monash hosts the East cluster of the Enterprise Grid plus an extra sixteen dual CPU dual core (64 cores) Intel Xeon 5160 @ 3Ghz based compute nodes with 8GB RAM and GigE interconnect. East and the other Enterprise Grid clusters are available for opportunistic use by researchers in areas other than computing, particularly in collaboration with the MeSsAGE Lab team. East adds approximately 2 Teraflops to the computational power at Monash.

Usage information:

Enterprisegrid Usage

Access Grid

MeSsAGE Lab has an Access Grid which is an ensemble of resources including multimedia large-format displays, presentation and interactive environments, and interfaces to Grid middleware and to visualization environments. These resources are used to support group-to-group interactions across the Grid.

Tile display wall for high resolution scientific visualisation

Monash OptiPortals

High Definition Video seminar room

HD Video

Expertise and experience

Our expertise and experience with various middleware, software systems, and standards is integral to our research capability and hence we include this human resource as part of the lab infrastructure. We like to share our experience, following are some miscellaneous items detailing experience and sharing tools that may be of use or interest to various communities.

• The Sun Grid Engine qsub helper developed for use on the Enterprise Grid clusters can add virtual queues and automated access resolution to SGE , this is particularly useful for jobs coming in off the grid (e.g. through Globus ) as existing interfaces do not allow for a full range of requests.
• Matlab compilation: We have noticed increasing demand for parameter exploration and Monte Carlo modeling over codes developed in Matlab. Software and license availability typically have a limiting factor on the level of parallelism possible though. Especially for grid runs we recommend users compile their Matlab code, this guide provides some instructions to get you started.