Enrollment Closed. Class limit reached. Please contact Bob Garza at garza@tacc.utexas.edu to get on a waiting list for this class.
The Sun Constellation Cluster, Ranger, is a computational resource available to the National Science Foundation research community, providing users access to a compute cluster with a theoretical peak performance of 579 Teraflops. A description of Ranger can be found at the following URL.
http://www.tacc.utexas.edu/resources/hpcsystems/#ranger
Instructors will present various topics covering the architecture and proper use of Ranger and supporting infrastructure. This class will be of particular interest to computational scientists interested in developing codes for and porting codes to Ranger. The course will include lectures, labs, and discussions on:
- Developing parallel-programming skills in OpenMP and MPI
- Understanding performance aspects of the AMD Architecture
- Using the SGE batch system
- Optimizing code
- Debugging code
- Compiling and running example codes
- Porting your own codes
Members of the Cornell University Center for Advanced Computing will be available during the entire training session to assist users in their code development and porting efforts.
Engineering Library, Carpenter Hall (Blue Room)
Cornell University
Ithaca, NY
Visitor Information, including maps, travel directions, and nearby accommodations, can be found at the following URLs:
- Cornell Maps - http://www.cornell.edu/maps/
- Carpenter Hall - http://www.library.cornell.edu/node/163
- Getting to Ithaca - http://www.cornell.edu/visiting/ithaca/visiting.cfm
Class attendees traveling to the training are responsible for arranging and paying for their own travel, daily expenses, and hotel accommodations.
Travel directions to Cornell University and nearby accommodations can be found at:
Visiting Cornell's Ithaca Campus: http://www.cornell.edu/visiting/ithaca/visiting.cfm
Visiting Ithaca, NY: http://visitithaca.com/lodging/search.cfm
| Time | May 28, 2009 | |
| General Topics | 9:00 | Welcome, Overview, and Hardware |
| 9:30 | HPC Environment | |
| 11:00 | Introduction to Parallel Programming | |
| 12:00 | Lunch | |
| MPI Programming | ||
| 1:00 | Introduction to MPI | |
| 2:00 | Lab: Introduction to MPI | |
| 3:00 | Programming with MPI | |
| 4:00 | Lab: Programming with MPI | |
| 5:00 | Adjourn |
| Time | May 29, 2009 | |
| Hybrid Programming | 9:00 | Programming with OpenMP |
| 9:30 | Lab | |
| 10:00 | Hybrid Programming with OpenMP and MPI | |
| 11:00 | Lab | |
| 12:00 | Lunch | |
| Code Improvement | ||
| 1:00 | Optimization and Scalability | |
| 2:00 | Lab: Optimization and Scalability | |
| 3:00 | Profiling and Debugging | |
| 4:00 | Lab: Profiling and Debugging | |
| 5:00 | Adjourn |