Phenomena

Exascale Simulations

Featured Video Play Icon

Capturing what goes on inside a combustion engine is incredibly difficult. It’s a problem that depends on turbulent flow, chemistry, heat transfer, and more. To represent all of those aspects in a numerical simulation requires enormous computational resources. It’s not simply the realm of a supercomputer; it requires some of the fastest supercomputers in existence.

Exascale computing, like that used for the simulations in this video, is defined as at least 10^18 (floating-point) operations per second. For comparison, my PC has a recent, high-end graphics card, and it’s about a million times slower than that. These are absolutely gigantic simulations. (Image and video credit: N. Wimer et al.)

3 comments
  1. Jonathan Bull

    Very impressive. I would like to know how these results helped with the piston design. Also, they said these are exascale-capable codes, not actual exascale simulations. Only 500 GPUs and a few billion DoFs is large scale but not quite exa I think.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: