Hank Hoffmann, assistant professor of computer science at the University of Chicago explains how he wants to tackle the power problem for Exascale systems in ACSR Discovery. According to the magazine, Hoffmann received a Department of Energy Early Career Research Program award this year to explore what he calls CALORIE. This ia a constraint language and optimizing runtime for exascale power management. In essence, Hoffmann envisions a technique that lets scientists focus on the application while CALORIE takes care of the power.
The article also explains, well tries to explain, how much power it takes to run an exascale computer: "The energy to power an exascale computer is comparable to the power needed to light 100-watt incandescent bulbs stacked end-to-end and climbing 2.3 times the height of Mt. Everest." Perhaps because I never climbed the Mount Everest, but 20 MW (Megawatt) sounds easier to me.
And do you power on the light bulbs? Do you then add the socket height? Or is it already added? And what about the power loss in the power cables all the way up the Mount Everest? And how do you get a 100 Watt light bulb? You cannot buy them in Europe anymore.
For simplicity let us say an Exascale uses as much power as about a million light bulbs.