Over at TechRadar, Julian Fielden from OCF writes that the users faced with almost insurmountable energy and cooling challenges will likely avoid owning and housing their own Exascale computing facilities and look to the “cloud” and on-demand services provided by much larger international suppliers.
In this episode of This Week in HPC, Addison Snell and Michael Feldman from Intersect360 Research discuss a recent presentation on Exascale by Jack Dongarra from the University of Tennessee. After that, Addison and Michael preview Karlheinz Meier’s ISC’14 keynote on the topic of “Brain-derived computing beyond Von Neumann – achievements and challenges.”
Philippe Ricoux from Total presented this talk at the Stanford HPC Conference. “The objective of this Support Action, co-funded by the European Commission is to build a european vision and roadmap to address the challenges of the new generation of massively parallel systems composed of millions of heterogeneous cores which will provide multi-Petaflop performances in the next few years and Exaflop performances in 2020.”
Scott Fadden from IBM presented this talk at the Stanford HPC Conference. “What does it mean to provide data to an Exascale system? Many believe that the current model of adding more disks and installing faster networks won’t get us there. So how do you get the right data to the right processor at the right time? How dow begin to leverage new storage technologies? This presentation explores some the Exascale challenges and provides insight on what is being done today to learn about and prepare for managing data in an Exascale system.”
Barry Bolding from Cray presented this talk at the 2014 HPCAC Stanford HPC & Exascale Conference. “Productive Exascale is not simply about achieving a set of technologies and performance metrics, it is about providing systems that fit into the production scientific workflow environments that will exist at the end of this decade.”
DK Panda from Ohio State University presented this talk at 2014 HPC Advisory Council Stanford Conference. “This talk will focus on programming models and their designs for upcoming exascale systems with millions of processors and accelerators. Current status and future trends of MPI and PGAS (UPC and OpenSHMEM) programming models will be presented.”
Intel’s Mark Seager presented this talk at 2014 HPC Advisory Council Stanford Conference. “In this talk, we will review the many challenges of building practical Exascale systems by the end of the decade and Extreme scale systems in the 2020s. Some of these challenges, such as extreme levels of parallelism, have direct impact on applications, while others, such as new data paradigms, offer real breakthrough application and scientific opportunities.”