In recent years, many researchers have debated on the technology advances that would enable the construction and deployment of Exascale computing platforms. Architectural challenges were identified, as were some requirements for a change in the system software. This included considerable discussion on whether applications would best be served by an evolutionary or a revolutionary exascale programming model.
Many of these discussions were held at an international scale. In the US, the Department of Energy has funded several modest efforts that have started to explore these challenges. More exascale research and development seems likely in the near future.
In this talk, we revisit the early discussions on programming models for exascale and look, in particular, at how the evolutionary approaches are evolving to meet the challenges of computation at this scale.
PH: 631-632-4629 | FX: 631-632-4125 | iacs@stonybrook.edu