April 18-19, 2012
Pittsburgh Supercomputing Center
300 S. Craig St., Pittsburgh, PA
PSC, together with NVIDIA and PGI, is pleased to announce a hands-on workshop on GPGPU programming using OpenACC, an open standard for compiler directives to make accelerators easier to program. Using OpenACC directives in standard C and Fortran, programmers insert compiler hints into their code to automatically execute compute-intensive regions of code on accelerators, simplifying code development and improving performance portability.
While GPGPU programming experience is not required, a working knowledge of C or Fortran is required. There will be ample time for hands-on experience, so participants are encouraged to bring their own codes. After two days, codes may experience significant acceleration with OpenACC.
- Welcome & Introduction to PSC's Computing Environment
- Overview of Parallel Computing
- Introduction to OpenACC Programming (with hands-on exercises)
- Using OpenACC with GPU Libraries
- GPU Parallel Execution Model / Architecture
- Performance Profiling and Tuning
- Advanced OpenACC Features
- Miscellaneous OpenACC Topics