Dr. Michael M. Resch of the University of Stuttgart has professorships, degrees, doctorates and honorary doctorates from around the world, he has studied and taught in Europe and the U.S., but for all the work he has done in supercomputing for the past three-plus decades, he boils down his years in HPC to working with the same, if always improving, von Neumann architecture. Hes eager for the next new thing: quantum. Going to quantum computing, we have to throw away everything and we have to start anew, he says. This is a great time.
In This Update. From The HPC User Forum Steering Committee
By Steve Conway and Thomas Gerard
After the global pandemic forced Hyperion Research to cancel the April 2020 HPC User Forum planned for Princeton, New Jersey, we decided to reach out to the HPC community in another way by publishing a series of interviews with members of the HPC User Forum Steering Committee. Our hope is that these seasoned leaders perspectives on HPCs past, present and future will be interesting and beneficial to others. To conduct the interviews, Hyperion Research engaged insideHPC Media.
We welcome comments and questions addressed to Steve Conway, sconway@hyperionres.com or Earl Joseph, ejoseph@hyperionres.com.
This interview is with Michael M. Resch. Prof. Dr. Dr. h.c. mult. He is dean of the faculty for energy-process and biotechnology of the University of Stuttgart, director of the High Performance Computing Center Stuttgart (HLRS), the Department for High Performance Computing, and the Information Center (IZUS), all at the University of Stuttgart, Germany. He was an invited plenary speaker at SC07. He chairs the board of the German Gauss Center for Supercomputing (GCS) and serves on the advisory councils for Triangle Venture Capital Group and several foundations. He is on the advisory board of the Paderborn Center for Parallel Computing (PC2). He holds a degree in technical mathematics from the Technical University of Graz, Austria and a Ph.D. in engineering from the University of Stuttgart. He was an assistant professor of computer science at the University of Houston and was awarded honorary doctorates by the National Technical University of Donezk (Ukraine) and the Russian Academy of Science.
He was interviewed by Dan Olds, HPC and big data consultant at Orionx.net.
The HPC User Forum was established in 1999 to promote the health of the global HPC industry and address issues of common concern to users. More than 75 HPC User Forum meetings have been held in the Americas, Europe and the Asia-Pacific region since the organizations founding in 2000.
Olds: Hello, Im Dan Olds on behalf of Hyperion Research and insideHPC, and today Im talking to Michael Resch, who is an honorable professor at the HPC Center in Stuttgart, Germany. How are you, Michael?
Resch: I am fine, Dan. Thanks.
Olds: Very nice to talk to you. I guess lets start at the beginning. How did you get involved in HPC in the first place?
Resch: That started when I was a math student and I was invited to work as a student research assistant and, by accident, that was roughly the month when a new supercomputer was coming into the Technical University of Graz. So, I put my hands on that machine and I never went away again.
Olds: You sort of made that machine yours, I guess?
Resch: We were only three users. There were three user groups and I was the most important user of my user group because I did all the programming.
Olds: Fantastic, thats a way to make yourself indispensable, isnt it?
Resch: In a sense.
Olds: So, can you kind of summarize your HPC background over the years?
Resch: I started doing blood flow simulations, so I at first looked into this very traditional Navier-Stokes equation that was driving HPC for a long time. Then I moved on to groundwater flow simulations pollution of groundwater, tunnel construction work, and everything until after like five years I moved to the University of Stuttgart, where I started to work with supercomputers, more focusing on the programming side, the performance side, than on the hardware side. This is sort of my background in terms of experience.
In terms of education, I studied a mixture of mathematics, computer science and economics, and then did a Ph.D. in engineering, which was convenient if youre working in Navier-Stokes equations. So, I try to bring all of these things together to make an impact in HPC.
Olds: What are some of the biggest changes youve seen in HPC over your career?
Resch: Well, the biggest change is probably that when I started, as I said, there were three user groups. These were outstanding experts in their field, but supercomputing was nothing for the rest of the university. Today, everybody is using HPC. Thats probably the biggest change, that we are moving from something where you had one big system and a few experts around that system, and you moved to a larger number of systems and tens of thousands of experts working with them.
Olds: And, so, the systems have to get bigger, of course.
Resch: Well, certainly, they have to get bigger. And they have to get, I would say, more usable. Thats another feature, that now things are more hidden from the user, which makes it easier to use them. But at the same time, it takes away some of the performance. There is this combination of hiding things away from the user and then the massive parallelism that we saw, and thats the second most important thing that I think we saw in the last three decades. That has made it much more difficult to get high sustained performance.
Olds: Where do you see HPC headed in the future? Is there anything that has you particularly excited or concerned?
Resch: [Laughs] Im always excited and concerned. Thats just normal. Thats what happens when you go into science and thats normal when you work with supercomputers. I see, basically, two things happening. The first thing is that people will merge everything that has to do with data and everything that has to do with simulation. I keep saying its data analytics, machine learning, artificial intelligence. Its sort of a development from raw data to very intelligent handling of data. And these data-intensive things start to merge with simulation, like we see people trying to understand what they did over the last 20 years by employing artificial intelligence to work its way through the data trying to find what we have already done and what should we do next, things like that.
The second thing that is exciting is quantum computing. Its exciting because its out of the ordinary, in a sense. You might say that over the last 32 years the only thing I did was work with improved technology and improved methods and improved algorithms or whatever, but I was still working in the same John von Neumann architecture concept. Going to quantum computing we have to throw away everything and we have to start anew. This is a great time. I keep telling my Ph.D. candidates, go for quantum computing. This is where you make an impact. This is where you have a wide-open field of things you can explore and this is what is going to make the job exciting for the next 10, 12, 15 years or so.
Olds: Thats fantastic and your enthusiasm for this really comes through. Your enthusiasm for HPC, for the new computing methods, and all that. And, thank you so much for taking the time.
Resch: It was a pleasure. Thank you.
Olds: Thank you, really appreciate it.
Here is the original post:
The Hyperion-insideHPC Interviews: Dr. Michael Resch Talks about the Leap from von Neumann: 'I Tell My PhD Candidates: Go for Quantum' - insideHPC
Read More..