Author_Institution :
Scalable Modeling & Anal. Syst., Sandia Nat. Labs., Livermore, CA, USA
Abstract :
As high-performance computing (HPC) advances toward exascale-class machines, there is increasing realization within the HPC community that major challenges lay ahead. These challenges span not only the system architecture and hardware, but also the entire software stack, including the applications. In particular, several key assumptions that are central to the current HPC application programming paradigm will not hold at this scale. In this talk we examine some of the key challenges from the perspective of the application programmer, and discuss ways to address or at least mitigate them. Hopefully, this will be somewhat of a ´call to arms´ for application developers who wish to prepare for the coming wave of exascale-class HPC systems.
Keywords :
computer architecture; mainframes; parallel machines; parallel programming; software architecture; HPC application programming paradigm; exascale-class HPC systems; exascale-class machine programming; hardware; high-performance computing; software stack; system architecture;
Conference_Titel :
Parallel and Distributed Processing Workshops and Phd Forum (IPDPSW), 2011 IEEE International Symposium on
Conference_Location :
Shanghai
Print_ISBN :
978-1-61284-425-1
DOI :
10.1109/IPDPS.2011.408