Steve takes us deep into the field of high performance computing and how AI is impacting it.
Sramana Mitra: Let’s start by having you introduce yourself a little bit as well as Cray’s activities in the domain of AI currently.
Steve Scott: I’m the Chief Technology Officer at Cray. I’ve been with Cray most of my career. I actually was a summer intern here for two years during grad school and have been here ever since except for a three-year walkabout I took a few years back. I went and spent two years as the CTO of the Tesla GPU computing Group at Nvidia and spent a year in the platforms group at Google and then came back in 2014.
Cray is all about high-performance computing. We’re commonly known as the supercomputing company. We were founded back in 1972 by Seymour Cray who’s largely considered as the father of supercomputing. We have been about building the fastest computational tools for several decades now to help governments and scientific leaders push the frontiers in a number of areas.
Our mission really is to build computational tools that help people change the world. Historically, we have not been focused on AI. Cray has been a supercomputing company.
Sramana Mitra: That’s the way I’ve known it in all these years.
Steve Scott: Exactly. People use our systems to solve their most challenging computational problems. This could involve the area of national security. Think cryptography and nuclear stewardship. Whether you’re in biology, physics, chemistry, everybody uses computational science to drive the discipline forward.
Computing has become the universal scientific instrument alongside theory and experiments because simulation can be used to model the world and ask questions about the world that you can’t get through direct observation or through experiment. All of the areas use computation.
Why are we here talking about artificial intelligence if Cray is a supercomputing company? That’s because we’re seeing artificial intelligence start to have a profound impact on what our customers do. The traditional methods of just using simulations and modeling are starting to open up and we’re finding ways to use artificial intelligence to help enhance or augment traditional simulations.
We’re pretty excited about it at Cray. We started looking at data analytics probably five years ago. In the past few years, the whole area of AI has really been taking root.
Sramana Mitra: I was going to ask you to highlight what are the big changes that are happening to your world because AI is becoming mainstream. Obviously, the speed of computation is a big part of the success of AI.
Today, some of the use cases that are really taking off are in cyber security and real-time bidding in the adtech world. These are split-second decisions that have to be made and that requires incredible computing power. What is all that doing to your world of supercomputing hardware?
Steve Scott: Cray is pursuing AI in two different ways. We’re first looking at the application of AI to high-performance computing and we’re also looking at the application of high-performance computing to AI. In terms of application of AI to high-performance computing, we are looking at mostly deep neural networks but other forms of AI as well. We’re finding ways where that they can substantially accelerate or improve traditional simulation.