Going to the Moon via the Cloud

Before the widespread availability of these types of computers, companies built expensive prototypes to test their designs. “We actually built a full-size prototype and ran it to the end of its life before putting it on site,” said Brandon Haugh, a core design engineer, of a nuclear reactor he had worked on with the U.S. Navy . “That was a 20-year multibillion dollar test.”

Today, Mr. Haugh is director of modeling and simulation at California nuclear start-up Kairos Power, where he refines the design of affordable and safe reactors that Kairos hopes will accelerate the global transition to clean energy.

Nuclear power has long been recognized as one of the best options for carbon-free power generation – aside from its prohibitive cost. However, Kairos Power’s advanced reactors are designed to produce electricity at a cost competitive with natural gas.

“The democratization of high-performance computing has now gone through to launch, so companies like us can iterate quickly and move from concept to field use in record time,” said Haugh.

However, high performance computing in the cloud has also created new challenges.

In recent years there have been a variety of custom computer chips designed specifically for certain types of math problems. Similarly, there are now different types of storage and network configurations in high performance computing. And the different cloud providers have different specializations. One may be better at calculating fluid dynamics while another is better at structural analysis.

So the challenge is to choose the right configuration and get the capacity when you need it – because demand has grown dramatically. Scientists and engineers are experts in their fields, but not necessarily in server configurations, processors, and the like.

This has led to a new kind of specialization – experts in high-performance cloud computing – and new cross-cloud platforms that act as one-stop shops where companies can choose the right combination of software and hardware. Rescale, which works closely with all major cloud providers, is the dominant company in this area. It combines computing problems for companies like Firefly and Kairos with the right cloud provider to provide computers that allow scientists and engineers to solve problems faster or at the lowest possible cost.

Comments are closed.