Communications of the ACM
Fixating on Exascale Performance Only Is a Bad Idea
Fixing on the exascale performance milestone misses an important point, Sadasivan Shankar told the recent HPC User Forum.
Credit: YouTube
A concentration on realizing exascale computing performance to the exclusion of all else will not address the grand societal and scientific challenges, according to Harvard University lecturer Sadasivan Shankar. He proposes an emerging materials science model called in silico inverse design, which involves distinct computational requirements that are different and challenging in comparison to direct design.
Shankar theorizes scientific and technological advancements have exceeded the capabilities of direct design, while the challenges confronting society and science share common technological disruptors — design, materials and devices, sensing, automation, and more — best resolved via inverse design. Shankar defines inverse design as "the ability to use predictive capabilities, to design a material, which when synthesized in the manufacturing line will exhibit targeted properties." He thinks the push to reach the exascale milestone is being played up while giving the practicality of developing real-world exascale applications little consideration.
Shankar says the uses and requirements of inverse design methodology ought to inform such initiatives. He believes materials science is particularly relevant to this notion, as the field directly impacts economies and progress. Shankar says semiconductor materials synthesis is one area that could benefit from inverse design.
From HPC Wire
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found