Carleton University - Canada’s Capital University Carleton University - Canada’s Capital University Sitemap
Contact SCS
Campus Map
Computer Science Search:
Powered by Google
News & Seminars Future Students Current Students SCS Research People Tech Support
SCS Research Groups
CU Image School of Computer Science Research Group
High-Performance Computing Virtual Laboratory
Professor Dr Jörg-Rüdiger Sack
Group Contact

Office: 5350 HP
Phone: 1-(613)-520-2600 x4361
Email: sackatsignscs.carleton.ca
There are many computationally intensive applications in which results must be delivered quickly. Such time-critical applications may pertain to weather prediction, search and rescue, emergency response time and military applications. Also, there are less critical applications that must process huge amounts of data at a time (such as a set of satellite images, each being 100MB). In many cases, the only solution to these problems lies in parallel computing. This area of research involves the use of multiple processors that work together on a problem to solve it within necessary time constraints. The input data is often distributed across multiple processors so as to share the workload. Despite its advantages, parallel and distributed computing carries with it various design and implementation challenges which make it more difficult to develop software. There are various issues dealing with data distribution, load balancing, deadlock, over processing, inter-processor communication speed, inter-processor topology, etc.

The strength of the School of Computer Science lies in both its expertise in parallel and distributed computation, and its access to state-of-the-art equipment. Aside from the various resident computing facilities, Carleton is an integral part of the High-Performance Computing Virtual Laboratory, or HPCVL (www.hpcvl.org). HPCVL was formed by a consortium of four Ontario universities, Carleton University, Queen's University, The Royal Military College of Canada, and the University of Ottawa, and now includes Ryerson University and Seneca College in Toronto. HPCVL is dedicated to providing researchers at member institutions and researchers from elsewhere in Canada with the secure High Performance Computing (HPC) resources they need to conduct innovative research in a broad spectrum of disciplines. To date, HPCVL has received over $73 million in support from the four participating universities, CFI, ORDCF, OIT, SUN Microsystems, IBM, and Entrust.

Carleton's HPCVL laboratory houses a 20-node SUN Microsystems SMP parallel computer, a terabyte server, and a Beowulf cluster. The cluster consists of 32 nodes with dual 1.7GHz Xeon Processors with 1.0GB RAM, and 32 nodes with dual 2.0GHz Xeon Processors with 1.5GB RAM per node, for a total of 128 CPUs. The cluster nodes are interconnected via a Cisco 6509 switch using Gigabit Ethernet, and 8 of the nodes are equipped with high-end graphics cards. In 2006, the laboratory will be undergoing significant architectural and equipment upgrades. The cluster will more than double in computational and storage capacity, providing even greater computing power and state-of-the-art resources.

Lab Number: HP5180
Group Web Page: www.hpcvl.org

Research:

Members: