Why the World is Starting to look like a Giant HPC Cluster

Print Friendly, PDF & Email

In this video from the Dell EMC HPC Community Meeting, Jimmy Pike from Dell EMC describes why the world is starting to look like a giant HPC cluster.

insideHPC: Jimmy, welcome. I wanted to ask you about an interesting point you made in your talk. You said that the world is really becoming like a big HPC cluster. What did you mean by that, Jimmy?

Jimmy Pike: It’s really not. But I think it’s a nice line to capture what’s going on. If you look at things, things are clustered together. I mean, if you go all the way back to, I think, the first cluster, I bet, was a Beowulf cluster in 1994 or something like that, but it was the beginning in many ways of putting machines together in parallel to make them do a lot more work than you can do with a single machine. And honestly, that’s what our entire industry has focused on for a long time, and I think that is the way we’re going to go in the future. So that’s what I mean by that.

insideHPC: And how does AI come into play here when we look at these traditional HPC in this?

Jimmy Pike: Well, so AI and HPC are very related. In fact, AI is not– I don’t like the term AI, machine learning, deep learning I like. But AI, machine learning, is not an HPC workload. However, it takes an HPC machine to do it. If you look at HPC, generally, you take a model or things like that, you turn it into an extraordinarily large amount of data, and then you go find some information for that data. Machine learning, on the other hand, takes an extraordinarily large amount of information and collapses it into an idea or a model. So it’s kind of the inverse in many ways, although that’s a pure solution, but I think you get the idea.

insideHPC: Yes. For inferencing?

Jimmy Pike: Yes. For inferencing. In fact, the scientific computing part is a very high precision floating point model where we do a lot of very accurate calculations. And machine learning, on the other hand, is a probabilistic study that says it might be this or it might be that.

insideHPC: So when you look at these machines, these devices, these setups, it sure looks like HPC to me from the outside.

Jimmy Pike: The machine certainly looks like HPC. Now, I think the future will be where we merge these two disciplines together, and you apply a machine learning things to either limit the amount of simulation you have to do with HPC, or you take machine learning to interpret the results that happen with HPC.


Jimmy Pike is a seasoned veteran with more than forty years of industry experience. He has served in key technical and executive roles at Dell, Intel, AT&T, NCR, and Harris as a pioneer in complex, multiprocessing, Hyper-scale and HPCC systems. He has a proven history of defining, developing, and delivering successful products to the industry and is noted for his “down-home”, practical manner. Jimmy has successfully managed geographically separated development organizations with management experience in both line and matrix organizations. He has a comprehensive knowledge of technology (hardware, software, and data center environments) and a keen understanding of areas like Cloud Computing and hyper-scale; HPCC; Data Center Architectures; Virtualization (server, server hosted clients, and clients); Hardware: Multi-processing – CPU/Memory/Cache & IO; Software: OS / Management / application and the emergence of IoT. He has more than 50 patents that are awarded or in process and the ability to move conversation back and forth between technical and management ranks at virtually any level.

​The Dell EMC HPC Community is a worldwide technical forum that fosters the exchange of ideas among researchers, computer scientists, technologists, and engineers and promotes the advancement of innovative, powerful HPC solutions. The next Dell HPC Community Meeting takes place at SC18 in Dallas.

Check out our insideHPC Events Calendar