The growing adoption of the affordable clusters and multicore systems for high-end research computation in recent years represents the dawn of an "HPC for the masses" era. However, simply buying and making the hardware available to everyone does not mean the new era has arrived. HPC is about exploiting the parallel computing capability to speed up computation. Therefore, making this process more user-friendly and the speedup easier to achieve are key to the effective adoption of the technology by the masses. At Computer Centre, we are working on a number of initiatives to enable more researchers to benefit from the use of HPC technologies:

• MyWinCluster Service - offers seamless access to HPC cluster from Windows desktop/laptop.
• Parallel File System - enables speedup in accessing large dataset during computation.
• NUS Grid - enables access to unprecedented amount of distributed parallel computing resources for solving large scale research problems.
• Parallel Programming Tools - provides parallel mathematical libraries and code optimisation tools to facilitate parallel application development.
• Off-the-shelf Parallel or Grid-enabled Application Software - allows instant use of parallel computing resources and capabilities.
• HPC Integrated Portal - (development in progress) provides one-stop HPC services at Computer Centre.
• Application Portals - (development in progress) such as the HPCBio Portal which provides user-friendly web-based interface to some parallel or Grid-enabled Life Sciences/Bioinformatics applications.

We will share the details of some of these efforts in this issue of HPC@NUS.

Wishing all a joyful and productive 2008!

The Matlab Distributed Computing Toolbox is a powerful tool that enables PC clients to access the computational nodes in MyWinCluster to run parallel Matlab computations. This article introduces the interactive parallel mode built into the DCT, which enables users to interactively run their command in parallel. For more information on Matlab Distributed Computing Toolbox, please contact Yeo Eng Hee at .

The Matlab Bioinformatics Toolbox comes with a rich set of examples and tutorials. We examine a demo to see how a distributed bioinformatics application is constructed. The distributed code calls functions in the Distributed Computing Toolbox (DCT). This Toolbox and the Distributed Computing Engine (DCE) are necessary components for developing and running parallel and distributed codes in Matlab. To find out more how these Toolboxes may be used in Matlab applications and how these applications may be run on SVU servers, please read on...

Image rendering can be very compute intensive and time-consuming. Distributed parallel computing is a very useful method to speed up the process. You might have heard of some studio factory making use of a computing farm to do some fantastic digital films. Read this article to understand how the Windows computing cluster system can be used to improve the rendering efficiency in the visualisation of engineering simulation results.

HPC is commonly applied in the financial sector within the areas of Treasury, Risk Management, Credit Card Fraud Analysis, Data Warehouse and others. It is very common for compute intensive work to happen in these areas. This article shares with you how you can run the Microsoft Excel based code on a Windows Computer Cluster Server to number crunch the massive data produced in the financial sector. For more information on how to run your Microsoft Excel based codes on SVU's MyWinCluster, please contact Yeo Eng Hee at .

Doing graphics modelling on your home PCs? Some users have already started doing this which was only made possible with the launch of MobileVislab, the remote and collaborative visualisation system. Previously when users worked at home, they could only work at the command line prompt and do text editing, or when they run graphical applications, the pictures moved slowly andonly provided jumping or jerky pictures. With MobileVislab, you will see the pictures moving smoothly and the refreshing rate of the pictures can be as high as 23 fps (frames per second), based on broad band speed of 1Mbps. Normally, a 10 fps frame rate is good enough to render a smoothly moving picture. The MobileVisLab brings all these graphics applications on your desktop: Abaqus, Analyze CFX, Ensight, Fluent, Gambit, GaussView, Matlab, Molden, VMD and ADT. Click here to find out more about it.