Towards Exascale Applications

The quest for Exascale is cruelly exposing key problems with our current approach to HPC. Applications cannot simply wait for the next generation of CPUs to enjoy improved performance, but need to deal with various forms of parallelism explicitly, which might require revised algorithms, appropriate programming environments, and a deep understanding of the hardware developments to come.

In this talk we review some of the key challenges exposed by the Exascale quest and present three KTH-led projects that deal with programming environments (EPiGRAM), algorithmic changes (ExaFLOW), and improved software and support (BioExcel).