What’s at stake? What impact will the U.S. presidential election have on the global HPC community and the race to exascale?
The immediate impact of the US presidential election is not about which party will win but rather that the system of responsible government is out to lunch. In spite of stated support for HPC in general and exascale in particular by the current administration, the U.S. is currently under spending international competition.
Neither party is committed to a sustained relationship with industry and academia, taking on the risk of real research so that American industry can increase its competitiveness through the end of this decade. Where required changes are incremental, as has been the case from the early 1990’s to the latter part of the last decade, this was a sound strategy with US investment principally in procurements of large systems and mission-critical application development. Industry knew the path to be followed with a little support for narrow places where focused innovation was needed (e.g., Lustre, early MPI).
But now the HPC community does not know the right path and industry cannot self-sponsor R&D in the presence of such uncertainty. Neither party appears to understand that the future of U.S. competency in many fields in science, technology, bioengineering and medicine, and industrial processes are all dependent on superiority in future generations of HPC. This requires partnership between government and industry.
The presidential election has left HPC research an orphan; it’s just not important to the electorate, to their representatives, and to the candidates. The solution to the economy challenge is to grow out of it and HPC is a critical tool to achieve this. HPC requires innovation through research, which must be government funded to enable future industry hardware and software products. That is what is at stake. The impact of the election is that HPC research has been diminished compared to that which is required strategically.
For related stories, visit The Exascale Report Archives.
