|Pragma based GPU programming with OpenACC|
|Karen Tomko, Ohio Supercomputing Center|
|Johnston Hall 338
October 11, 2012 - 12:30 pm
In this talk I’ll provide a brief introduction to the OpenACC standard. I’ll discuss current compiler support for the standard and present some examples of OpenACC constructs. Additionally I’ll discuss my experience using an early version of PGI’s OpenACC implementation for accelerating some loops from the measurement calculations of a Quantum Monte Carlo Condensed Matter Code.
Dr. Tomko, a research scientist at the Ohio Supercomputer Center, has been collaborating with computational scientists for 20 years. Her experience ranges from optimization and parallelization of crashworthiness simulation to development of a run-time adaptive parallelization scheme for wireless communications simulations. Current collaborations focus on Multi-scale Many Body methods, MPI communication, and Analysis frameworks for High Energy Physics. She is involved in curriculum development for the Ralph Regula School of Computational Science and the XSEDE Education Program and has taught distance courses on topics in high performance computing.