Learn Parallel Processing with CTC's Windows-Based HPC Workshop

The Cornell Theory Center (CTC) is offering a hands-on, two-day technical training course for members of organizations interested in implementing high-performance computing on Windows-based clusters. The workshop is scheduled for February 8-9, 2005, at 55 Broad Street, New York, N.Y. Representatives from all industries and sectors, including corporations, universities, and government agencies, are invited to attend. Space is limited. For more information or to register, visit www.tc.cornell.edu/services/edu/events/winhpc/ . The workshop will focus on technical training, with discussions about installing clusters and writing parallel applications. Specific sessions will include overviews of Windows-based HPC, cluster design, parallel programming, benchmarking, and hardware and software considerations. Attendees will also have an opportunity to build a Windows-based cluster. The course format includes lectures, demonstrations, and hands-on exercises. Specifically, the Windows High-Performance Computing Technical Training workshop will offer the following sessions: • Overview of CTC Solutions Center, High-Performance Computing • Designing and Monitoring an HPC Cluster • Parallel Programming with Message Passing Interface (MPI) • Benchmarking • Porting Applications • Building a Cluster “The model covered in this workshop reflects traditional message passing-based supercomputing where processes run on multiple computers and need to exchange information to complete a computation,” said CTC Chief Technical Officer David Lifka. “This model, which uses dedicated clustered servers, is ideal for computationally intense calculations that require all of the system’s resources (CPU, memory, network, bandwidth) and may take hours to complete.” CTC will be offering a new workshop, scheduled for February 10-11, 2005, also at 55 Broad Street, New York, New York, that offers an introduction to .NET and Web Services. This new workshop is designed to assist those who typically use clusters for many small calculations that take a few minutes or less for each computation. The computers can be shared by multiple computations and users. Using the .NET framework and Web Services typically allows the user to reduce the size of the code, which often results in fewer bugs. In addition, security breeches are fewer and optimization is better in the .NET framework. For more information or to register for either of the workshops, visit http://www.tc.cornell.edu/services/edu/events/winhpc/. For questions concerning registration, hotel, or travel, please contact Judy Miller at 212.363.2915 or judym@tc.cornell.edu. All other questions, including workshop content, may be directed to Susan Mehringer at 607.254.8777 or susan@tc.cornell.edu.