The Golden Energy Computing Organization (GECO) invites scientists and engineers to attend two up coming workshops. The first will be an An Introduction to High Performance Computing with RA - August 17-21. The second is An Introduction to Graphics Processor Programming for HPC - August 24. This form can be used to sign up for either of the workshops.

Please use this form to request to attend these workshops. Emailed requests to attend the classes will not be accepted.

An Introduction to High Performance Computing with RA - August 17-21

RA is a Linux cluster composed of 2144 Intel cores (268 nodes of 8 cores each), an Infiniband fabric, and 300 TB of disk space. The machine is rated at 23 TF (17 TF sustained). The operating system is Red Hat, file system management is Lustre, and job scheduling and management is carried out via Moab (PBS). The intended use of the machine is for large, parallel processing jobs using the message passing interface (MPI). Both Intel and PGI compiler suites are supported with PGI and DDT parallel debuggers.

This workshop will cover such topics as: Introduction to Parallel Computing, Running programs on RA, Basic to advanced MPI programming, batch scripts, debugging, OpenMP, and MPI IO. Materials from last year's workshop can be found at http://geco.mines.edu/workshop/class2. We are planning on offering this workshop in mornings. If enough people sign up we will do a repeat session in the afternoon.

GECO supports a website that includes an extensive help section to get new users started: http://geco.mines.edu/guide.

Introduction to Graphics Processor Programming for HPC - August 24.

Special purpose computing accelerators, such as graphic processing units, (GPU) have become an important platform for parallel high performance scientific computing. GECO recently took delivery of and installed a Nvidia Tesla S1070 system. This 1U rack mountable system contains 4 of the Nvidia Quadro FX 5600 GPU cards and has a peak computing performance of 4 trillion floating point operations per second, or 4 Teraflops. This new Tesla system is front-ended by one of the RA nodes. Each of the 4 graphics processing units (GPU) on the Tesla has 240 processing cores and 4 Gbytes of memory for a total of 960 cores and 16 Gbytes. The individual GPUs are connected to the front-end node via a PCI connector.

This workshop will cover such topics as: Overview of the Tesla GPU architecture, Threads programming, Cuda, Programming model, Performance Improvement, and library routines. This will be an all day workshop. Information on using this resource will be posted at http://geco.mines.edu/tesla as it becomes available.This system was paid for by using money requested as part of the Tech Fee proposal process. Tech Fee money is earmarked for equipment and infrastructure for student use. Therefore faculty will not be allowed to attend this workshop.

For questions please contact Dr. Timothy Kaiser at tkaiser@mines.edu

Emailed requests to to attend these classes will not be accepted. Please use this form.