Seminar Details:

LANS Informal Seminar
"On Convenience and Inconvenience in GPU Computing"

DATE: June 26, 2013

TIME: 15:00:00 - 16:00:00
SPEAKER: Karl Rupp, Postdoc, MCS, Argonne National Laboratory
LOCATION: Building 240, 1404-1405, Argonne National Laboratory

The desire for solving larger problems at shorter execution times has led to the use of graphics processing units (GPUs) in the scientific community. They promise to provide a large number of floating point operations per seconds as well as higher memory bandwidth than traditional configurations. Recently Intel also introduced their first accelerator platform, addressing the demand for hardware which is able to deal well with compute-intensive algorithms. The focus of this talk is to share experiences both for 'hard facts' such as mapping and tuning algorithms to GPUs as well as for 'soft facts' such as programming models which work for non-experts and for hardware from more than one vendor.


Please send questions or suggestions to Jeffrey Larson: jmlarson at anl dot gov.