Making Parallel Programming Synonymous with Programming for Linear Algebra Libraries
June 19, 2008
Robert van de Geijn (Professor of Computer Sciences, UT-Austin)
Abstract
We have invested heavily in hardware development but software tools and methods to use the hardware continue to fall behind. The Sky is Falling. Chicken Little has been proclaiming this now every time a new architectural paradigm shift has occurred. Panic is rampant. Money must be spent.
We respectfully disagree for the domain of dense linear algebra, which is a typical example used to motivate high performance. Over the last ten years, we have been developing software tools and methods targeting this domain specifically to stay ahead of architectural development. In this talk, I overview these methods and software tools, developed as part of the FLAME project. We show how, when applied to a broad range of architectures, including GPUs, they provide an out-of-the-box solution that attains high performance almost effortlessly.
Prof. van de Geijn received his Ph.D. in Applied Mathematics from the University of Maryland in 1987 and has been a faculty member at UT-Austin since 1987. He has contributed in the area of high-performance computing through his research on parallel computing, derivation of algorithms, and software architecture.
Download: Slide Presentation (pdf)
Host: Richard Barrett (rbarrett@ornl.gov), 241-1512