CIQ
HomeCIQ Webinars

HPC with Heart: Tapping into Your “Why”

May 11, 2023

This webinar will explore the “HPC with Heart” concept and how understanding your "why" can help you excel in your HPC endeavors. We'll discuss the importance of purpose-driven computing and how to uncover your own personal narrative to drive innovation and success.

You'll learn practical strategies for aligning your computing work with your values and overcoming common challenges and obstacles that can hinder productivity and fulfillment. Whether you're a seasoned HPC professional or just starting out in the field, this webinar will provide you with valuable insights and actionable tips for achieving more meaningful and impactful computing work.

Webinar Synopsis:

  • Introduction To Roundtable Topic

  • Rose’s Why

  • Introduction Of Panel

  • John’s Why

  • Dave’s Why

  • Alan’s Why

  • Brian’s Why

  • Forrest’s Why

  • How Do Personal Goals Influence HPC

  • Experiences That Enriched HPC Design

  • HPC Stepping Stones That Enrich Humanity

  • The Scientific Method Should Be Humbling

  • How Does HPC Improve Society

  • Real World Examples Of Personal Motivations In Computing Initiatives

  • Modeling And Simulation Application Use In Experimentation

  • Zane’s Why

Speakers:

  • Zane Hamilton, VP of Solutions Engineering, CIQ

  • Rose Stein, Solution Engineer, CIQ

  • John Hanks, HPC Principle Engineer, CZBioHub

  • Dave Godlove, Solutions Architect, CIQ

  • Alan Sill, Managing Director, Texas Tech/NSF CAC

  • Brian Phan, Solutions Architect, CIQ

  • Forrest Burt, HPC Systems Engineer, CIQ

Note: This transcript was created using speech recognition software. While it has been reviewed by human transcribers, it may contain errors.

Full Webinar Transcript:

Narrator:

Good morning, good afternoon, and good evening wherever you are. Thank you for joining at CIQ. We're focused on powering the next generation of software infrastructure, leveraging the capabilities of cloud, hyperscale and HPC. From research to the enterprise, our customers rely on us for the ultimate Rocky Linux, Warewulf, Apptainer support escalation. We provide deep development capabilities and solutions, all delivered in the collaborative spirit of open source.

Zane Hamilton:

Hello everyone, and welcome back to another CIQ webinar. How are you, Rose?

Rose Stein:

Awesome. So good to be here and see you, Mr. Zane and everybody out there. Hello. Good morning.

Zane Hamilton:

Absolutely. Thanks everyone for joining. We're back to another research computing roundtable. What are we talking about today, Rose?

Introduction To Roundtable Topic [5:42]

Rose Stein:

You know, this is actually a really interesting topic. We're talking about HPC with heart and how to really tap into your why, which is not a topic I think that talks about a lot, but it brings in the aspect of inspiration and purpose and that makes this kind of work. It's not just fun. There's an energy behind that when you have a purpose and what it is that you are doing, which totally makes sense because I don't know why anybody in their right mind, if they don't have that kind of purpose behind them, would spend the kind of time that HPC admins and developers and people in that world spend on creating really amazing open source, often, products for us to use.

Zane Hamilton:

Very true. That is very cool. So I know that while you and I are not HPC admins, what is your why?

Rose Stein:

Oh, that's deep. I feel a little bit on the spot.

Zane Hamilton:

I totally put you on the spot. What is your why for HPC? Why are you here

Rose’s Why [6:56]

Rose Stein:

Yeah, I mean, it's probably similar to what a lot of the guys are going to be talking about. And as with all things that are great and powerful, they can be used in a weird way, going in a direction that maybe does not inspire me and or can be used to go in a direction that is like, wow, that's really cool. So for any of us that have had any kind of health issue that has been resolved or solved, if there's any kind of medication or anytime that we are coming together to take data and information and understand it in a way that then we can put things into action, whether it's a medication or it's a service or it's making some changes in a way where we understand things at a deeper level at the human brain, maybe wouldn't be able to figure out on their own.

I think that's really what it is for me, is the opportunity and the possibility to do really great things. And I'm being kind of vague because it's so big, right? I mean, it could be anywhere, medicine, you know, to anywhere where we're like tapping into data. So it is really exciting to be a part of this. And when I talk to people that are doing research, like cancer research and things like that, and there's any way that we could help support them and make their life easier and better and simpler and get the results faster to save lives. That is really inspiring when we have those types of conversations with customers.

Zane Hamilton:

Absolutely. It's fantastic. Thank you for being put on the spot and sharing.

Rose Stein:

And back at you buddy. Back at you.

Zane Hamilton:

Well, let's bring in our panel and let's find out some more whys.

Rose Stein:

I am not letting this one go, Zane. You're going to give your why at some point.

Zane Hamilton:

I'll give my why. Absolutely. Welcome everyone. I appreciate you all joining. Let's go ahead and start with Griznog, introduce yourself again. It's been, it's been a couple weeks.

Introduction Of Panel [9:08]

John Hanks:

Hi. Yeah, I haven't been out here in a while. Yeah. I'm John. I'm a HPC principal engineer currently working at Chan Zuckerberg Biohub and have been doing HPC for a very long time.

Zane Hamilton:

Thank you, John. I've heard some of the story from you and I I love to get into that. Dave Godlove. Welcome.

Dave Godlove:

Hey, I'm going to say my name is Dave Godlove because there's so many Daves running around. So I'm one of the many Daves part of the Dave collective. I'm Dave Godlove. My background is primary science, so I used to be a neuroscientist working at the NIH. Then I got into HPC. I've been with CIQ for about a year now, and I've been kind of around the Apptainer community and working with Apptainer for quite some time.

Zane Hamilton:

Thank you Dave. Dr. Sill, welcome back. Good to see you.

Alan Sill:

Alan Sill. I've run the high performance computing center at Texas Tech University and one of the co-directors of a multi university industry university cooperative research center. I got into this from particle physics through the mechanism of being among the group of people who kept arguing about it after the meeting was over. And so those are the people who ended up doing it. So that's led to a career.

Zane Hamilton:

Absolutely. Thank you Alan. Brian Phan, welcome back. Good to see you.

Brian Phan:

Hey, good to be back. Brian Phan here. I'm a solutions architect here at CIQ. My background's in HPC administration and architecture. I got into HPC through the genetics and aerospace industry verticals.

Zane Hamilton:

Thank you Brian. Forrest, welcome back. Good to see you.

Forrest Burt:

Thank you, Zane. Good to see you as well. Good morning everyone. My name is Forrest Burt. I'm an HPC systems engineer here at CIQ. I got into HPC as a student while working as a student sysadmin on the high performance computing architecture at the university I was going to school at. I was working in the academic national lab HPC space. And, now I'm here at CIQ and enjoying the work we do here.

Zane Hamilton:

Thank you Forrest. So I'm going to go back to John. I'll go back up in the top corner. I want to hear everybody's why, so I want to go around and get everybody's why they're in HPC. I think we understand pretty much where we are, but why are you into it? And then we can dive in from that and know there's going to be other questions and kind of weave that way. John, why?

John’s Why [11:30]

John Hanks:

Not necessarily specific just to HPC but in general, I am well aware that someday I'm going to die. And when it happens, I would like to think that the world was better because I was here than worse because I was here. And that's kind of the motivation that controls or drives what I do when running a cluster and when doing HPC and it really shows up in the jobs that I've had along the way. I've always gravitated toward academia or medical research or something that I felt like was going to improve lives. I don't think I would've ever been a good fit or interested in doing HPC at a hedge fund, for instance. That's just not something that I would view as improving the world. So that's not something that would've motivated me to actually get up and show up for work every day.

Zane Hamilton:

That's great. Thanks John. And we'll come back to more questions on that in a minute. But Dr. Godlove?

Dave’s Why [12:29]

Dave Godlove:

Yeah, I mean, I couldn't have put it any better. The statement that someday you're going to die and you want to have a legacy, right? You want to leave something behind that has improved the world and made the world a better place because you existed. That has been something that has driven my entire career since I first got my first job pretty much. I will say that I feel that there's a little bit of a mixture there, because for me there's a mix of number one, wanting to leave my mark and make the world a better place through my work. But number two, there's also a little bit of a selfish part. I'm curious about stuff and I want to learn stuff, and I've got this itch of curiosity that I want to scratch.

But I think we could zoom out and we could wax really philosophical about human beings as a species and how that curiosity drives us as a species forward and actually serves humanity by helping us to discover things, which is our specialization within evolution. It is discovering stuff and using it and amassing knowledge and whatever. But, I'll leave that for later on in the discussion. But yeah, that's basically it. It's a mix of wanting to make the world a better place, and I'm curious about stuff and I want to learn more about it.

Zane Hamilton:

Thank you Dave. Alan?

Alan’s Why [14:04]

Alan Sill:

Yes. It's very hard to do better than John did. It was a high bar that you set for us there. I guess I just think of it in more personal terms just going into it. I think as Dave said you're driven by curiosity. Initially, I was curious about science. I was curious about it. I just thought that was a natural way to do things and keep exploring. Somewhere along the way, I found that things that were hard for other people in science or in especially computing were not hard for me. Now that's not to brag because I still can't play the piano or I already suck at the guitar and stuff. But, there's lots of things that almost everything other people are better at than I, but there were things that I could do that I couldn't understand why they were hard for other people.

When I went into grad school, I had a very ambitious goal to try to do something in science that nobody had done before. It was a grand vision that drove me. I wanted to try to push myself beyond some boundary. And, years later I realized that I had done that many times. That just partly by the company I kept and the things I managed to get involved with we found the top cork. We helped discover the Higgs. There were lots of things that would certainly check that off. So, somewhere probably 15 or so years ago, it changed me to be helping other people have that same feeling. And now almost a hundred percent of my time goes into that one way or another. Little asterisk on it for me is that I'm really interested in pushing some advances in the area of climate change. An area I can affect is in data centers. So a lot of my work goes into that. So these are all personal anecdotes, but I think perhaps I touched on some of the things that John and Dave raised.

Zane Hamilton:

Absolutely. So whenever you're talking about you were setting goals or you wanted to do something you aspire to do something, did you have a specific goal? Or was it just that I wanted to excel at something? So you left it generic enough that you could target a lot of things, or was there one specific piece that you were trying to solve?

Alan Sill:

Yeah, I knew what kind of science I wanted to do, and I didn't know exactly what so I didn't say I will discover that. I wanted to push myself to that level of excellence that I could help advance the field. And, then it was a matter of learning how to do that. And I was very, very fortunate. The little research group I joined at American University had a team at Stanford Accelerator. And, I was exposed to some very high level folks who would walk around in Birkenstocks and shorts. And it is a very nice environment and I just had the opportunity to learn from some very talented people. It worked out for me. Now, there's a lot of people it doesn't work out for.

I want to be clear that there's nothing guaranteed in science. That to get these kinds of rewards and to some degree the goals shouldn't be to get these kinds of rewards. It should be something you enjoy doing, right? You may never get acknowledged for the stuff you do. Nobody ever gets acknowledged as much as they want. So let me acknowledge all of you. You're all wonderful. The audience too, right? Every once in a while things do work out and you have a chance to say, I actually got to that mountaintop and then you have to decide what to do next. And for me, the last 15 or 20 years have been trying to cultivate the largest number of people to do the greatest number of things that they can do. And this really fits well with running a high performance computing cluster at university. Because every time we've talked about this before, every time you answer the phone or pick up a zoom call or look at a trouble ticket, there's somebody that needs your help. So, to some degree you guys at CIQ must have that feeling too, that you're doing this for a reason and then you're helping folks.

Zane Hamilton:

Absolutely. Thank you for that, Alan. Brian, what's your why?

Brian’s Why [18:44]

Brian Phan:

My why's pretty personal. I'm down to share this with everyone. My why is motivated by my parents. My parents were refugees from Vietnam. They escaped communist Vietnam in the seventies. They moved to North America. And, I think the reason why I do what I do is because growing up they've sacrificed a lot for me to have everything that I have today. So to really honor their sacrifice and everything I do my best to try to make a meaningful impact in this world. And I think high performance computing is an industry where I am able to do that and make these impacts beyond anything they could have ever comprehended or imagined. So yeah, that's a little bit of why I do what I do.

Zane Hamilton:

That's fantastic, Brian. Thank you.

Rose Stein:

Yeah, thanks for sharing that.

Zane Hamilton:

Forrest.

Forrest Burt:

Forrest’s Why [19:42]

Yeah. So my why in high performance computing, computing has always been a passion for me. I've been involved with the cloud and just with computing systems for a long time. Since I was little, I've really enjoyed the game interface with just the digital society that we have. Once I found out the scale that some people do it in HPC I was instantly hooked. I knew that just that level of system is what I wanted to see in action. I found along the way that while pursuing my passion in computing, it just so happens that high performance computing supports a lot of other people's passions in research or science or engineering, that type of thing. I found that it's very fulfilling to me personally to get to take the skills that I have and so directly apply them to the, oftentimes, life's work of other people and see how that allows them to then go and pursue their passions so massively on a different scale as HPC frequently provides.

I love computing. I'm very passionate about it. Once I got into HPC, I've always been someone who likes to teach people what I know and share the information that I'm aware of. Once I got into HPC, and just found that I'm somewhat decent explaining these technical concepts, or at least I'd like to think so, I found that it was incredibly fulfilling, as I said, to get to take people that have never used these systems and show them here's the incredible power that it has for you, and here's how you can harness it. Yeah, that's why I'm in HPC.

Zane Hamilton:

Thank you Forrest.

Rose Stein:

Forrest, you are definitely amazing at explaining, I mean, all of you guys are, but yes, an accurate self-assessment.

Zane Hamilton:

It's fantastic. I know everybody had other things they wanted to talk about. John, whenever you're looking at building a new cluster and that personal goal of yours, how does that influence the decisions you're making or the way that you're going about building a cluster or, high throughput computing? How does that impact the way that you are doing that?

How Do Personal Goals Influence HPC [21:59]

John Hanks:

For me specifically, we have, me and some of the people that I've worked with for a long time, we have a saying, if it doesn't support the science, there's no point in doing it. And so that's our mantra when we're designing a cluster, building an environment, whatever, everything is focused on how do you get the most science out of this pile of stuff for the least amount of money invested as possible. So that is the driving force. That's probably an offshoot of my choice of always working in academic environments because academic environments typically are resource constrained. You don't have a lot of money to waste on stuff. So you really have to be careful and focus what you do specifically on the problem being solved, which is not always easy when you're doing a general purpose cluster, but you do your best, which means a lot of communication with researchers, a lot of talking with users. I try to approach everything as if I am the grad student in the corner of the lab who knows all about computers and is running the cluster. So I consider myself a peer of the people that I support and a colleague, not a customer vendor relationship, which I think is antagonistic. So I try to avoid those as much as possible.

Zane Hamilton:

Thank you, John. Dave, you paused a little bit earlier as well, so I know there's more to this whenever you're looking at doing design utilizing HPC from what you've been through and where you've come from. So I'll let you dive into that a little more.

Experiences That Enriched HPC Design [23:35]

Dave Godlove:

Yeah, there's so much to say on this topic of having purpose and talking about your why. I come from HPC from using it as a scientist. And I go back to I'm always thinking about HPC from the perspective of a scientist and what's going to make a scientist's life more easy. How are we going to get more science out of it and stuff like that. I guess one thing that I can add to this is that I sort of approach science a little bit differently too than I think a lot of people do. I used to be a primary research scientist.

And what I used to do is I used to make neural recordings. I used to record electrical activity from individual neurons and from collections of neurons of animals as they were doing tasks to learn how we learn and how we get better at stuff. How we process reward and how we detect and process and learn from errors. Error monitoring was a big part of my research. And so I would a lot of times when I was a grad student and a postdoc doing this research, I'd be talking to a friend or an acquaintance or a family member or something explaining to them what I would do. I would always get to a point and there would be a question, so what's the practical application of what it is that you're doing right now?

And that, I think, is a question that makes a lot of scientists crazy. At one point in time, I used to be sort of hand wavy and be well, you know, there's some research which is involved. I mean, so we've gotten from this research some treatments for Parkinson's disease, which include deep brain stimulation and things like this. We've gotten some applications for people that want to move artificial limbs or want to move cursors on a computer that have syndromes, which prevents them from moving their arms and things. But the truth of the matter was my research was not applied. And those were the types of things that had come out of the types of research that I was doing.

I wasn't setting out to make those types of discoveries and my peers weren't either. And so I got to a point where I would react to that question and there would be no application for what I'm doing. And I wouldn't apologize for that because I would say as a species, as human beings, our specialization on the planet is to gather knowledge. And that's how we have excelled. And you could have, 50 years ago, you could have asked somebody so we're trying to put people on the moon, right? What's the practical application of that? And in fact, people did that. People asked what's the practical application of putting people on the moon? And I don't think that the scientists who were involved in that project would be well, there's going to be these cell phones that everybody's going to have in their pockets and they're going to have to have satellites circling the earth to work properly, and we're going to be able to disseminate human knowledge.

Really? No, it wasn't like that. It was just because it's a worthwhile thing to do. It's a worthwhile endeavor and we'll get practical use out of it that we don't see right now if we just pursue that. And so I think a lot of science is like that. And because I've got that mentality bringing that back to HPC when I'm helping scientists fulfill what they're doing, I know that that work is going to enrich our world and is going to make people's lives better regardless of whether it's got a real practical implication today. And some of it does when we're talking about modeling car crashes and saving car companies from having to crash real cars and allowing them to crash multiple cars and gather multiple data and stuff like that. And we're talking about things like that. Obviously, it's got real world practical implications, but a lot of it doesn't. And that's okay. And it still helps humanity and helps make the world a better place. I could go on and on. I'm not going to.

HPC Stepping Stones That Enrich Humanity [28:14]

Zane Hamilton:

I was fortunate enough to spend a couple of days with Dave recently, and we got into this over dinner a couple different times and went pretty deep on what he had done. And it is absolutely fascinating. It's to a whole level that I would've never dreamed. And it's amazing. So I appreciate you sharing with me.

Dave Godlove:

Thank you, Zane.

Zane Hamilton:

Absolutely.

Rose Stein:

Yeah, Dave, you brought up about Parkinson's, and so I have an interesting story and I'd like your guys' take on this as well. So about, oh, I don't know, like 12, 13 years ago, my son had some health issues and he had been on a lot of different antibiotics, when he was little and he had eczema and just all of these things happening ear infections, and that's why he was on the antibiotics and all this stuff. And, I'm a little bit more of like I don't know, people call me "hu ji woo ji" maybe, you know what I mean? Like, oh, you're just from California. I like Birkenstocks, Alan, I'm into it. I like intuition. I like to try things that are not scientific. That's my jam.

Zane Hamilton:

We're turning that into a short, by the way.

Rose Stein:

Totally. And so I started giving my son probiotics when he was very little and his eczema started to clear up. I told his pediatrician who also happened to be my stepfather. I told him, I was like, Hey, just so you know this because I tried all the creams and all the other stuff that they told me to do. This is helping. And he just shook his head and he's like, Rose, please, the gut bacteria doesn't have anything to do with the skin. That's ridiculous. Those were his words, right? I mean, this was a long time ago. Very smart man. He was just trained at a different time. Well, of course time goes by and amazing scientists like you guys and HPC admins allow people to have this incredible research. Now we actually know, oh, hey, that's a thing, right?

The gut bacteria actually does have a lot to do with the skin and your mental health and your emotional health and all of these things. Now, it's scientifically proven and we understand it a lot more. So he ended up getting Parkinson's along the way around that same time he was diagnosed. And I remember he called me up one day and he is like, well, Rose, it looks like you might be right. There's something about Parkinson's that part of it starts in the gut apparently, and it travels up the vagus nerve and does weird things in the brain. And so he was explaining this to me, and he's like, there's so much amazing science now. And he did all that brain stimulation stuff, and he got the little electrode thing implanted into his brain. He did everything, right?

He was just so into it, loving the science, wanting to be a part of all the trials. It was really cool and fun. That is one of those things where it's a real world example, right? Of science coming in and proving things that we might have an intuition about or a thought about or a question about. And really being able to come in and do the research and analyze the data and then produce really cool effects in the world. And so I I heard what you said, Dave, you were, I don't want to have to make that connection because science is just amazing on its own. And I would really love to know from each of you if there's anything like the car crash simulations or like Parkinson's research that you are aware of that some of the science that you've been a part of or helped enable what those things were.

The Scientific Method Should Be Humbling [31:59]

Dave Godlove:

I will jump in real quick and just say, hopefully quickly that, science, if you actually practice science, one of the things that it does is it makes you very humble. And it's just because it shows you how little you actually understand about the natural world. You set up an experiment, you read the literature, you say, okay, this is the next logical step in the experiment. I'm going to set this up so that if it ends with A, it's going to tell me something and if it ends with B, it's going to tell me something else. And then you do the experiment and you get X and you're like that doesn't make any sense at all. That was totally different. What? And anything you know? And then you go through that process over and over again until you become really humiliated. I've read all this literature. I came with all these preconceived notions and then you realize that the natural world is really complicated and really messy, and all the stuff that you think you know about it, a lot of it is really wrong. And that's something that science teaches you and it teaches you to be very, very humble.

Alan Sill:

Well, along these lines, let's be clear. It's not all good. We can all easily get into a point of a view where we try to tout the great virtues of science. But the problem is that these outcomes are used by humans and not always the way that the scientists intended. A really obvious example is the worldwide web. Is that a universal good? It was particle physicists that came up with it. I've always felt as a particle physicist, and it's unfair when these topics come up, because we have a huge number of consequential impacts on society. You know microwave ovens were invented by people trying to make accelerators. Well, the microwave cavity itself. There was an accelerator next to where I got my PhD thesis that they had built in a parking lot out of money they scrounged out of the lab budget because they couldn't get the energy research and development administration to fund it. It later became DOE. That accelerator discovered the charm. Cork won the Nobel Prize, and then within a few years 100% of the x-rays that came out of that accelerator because they didn't have much money, they built it small and a lot of the energy leaks out of the outside gets called syncreon radiation. Dozens of beamlines were built around that accelerator to use the syncreon radiation to make advances in biochemistry and smaller chips and stuff. So a hundred percent of this accelerator that had been built for pure research purposes has ended up using applications and the whole field of sync on radiation. They built all accelerators. So it's just, I could go on like this for a very long time. Like Dave, the positive stories from pursuing research for its own benefit are huge.

But let's be fair, there are negative outcomes because these technologies land and someone says, I can make Bitcoin out of that. And then heats the planet up by three degrees right there. So it's humans that use this. So it's up to us to pick our directions and constantly be aware every moment that we have choices to make about how to spend our, for me, that means I spend a lot of my time trying to figure out how to mitigate the impact on the environment of data centers. And I think it's possible. But who's to say that something we come up with might not be used for some bad purpose?

Dave Godlove:

Hmm, that's true.

John Hanks:

We're talking about science here and there's a couple of things that I have that are pet peeves about people's view of science. Most people view science as if it's just another religion. I'm not talking about the people who are actually doing science, I am talking about the general population. They view it as if science is just another religion. And it absolutely is not a belief. Science is literally a set of steps you take to overcome bias and belief. It's as far opposite of what people normally think of science as you could possibly get. And it just drives me crazy when I hear people talk about science as if, oh, the science is in. No, the science is never in. That's the definition of science is that it's never in. There's nothing more solid on a scientific foundation than the theory of evolution. We still call it a theory. And if new information comes up, we'll throw it out and go with the new information that disproved it, right? Yeah. That's how science works.

Alan Sill:

I agree with you. And even your starting points. People think you sit down and come up with very rational starting points. The starting point doesn't matter. You can have all the biases you want coming in. It's what you do next that makes it science or not. If you perform an experiment or pursue a line of reasoning that has one outcome, if idea A is right and another outcome, if idea B is right, and if you perform that test that's doing science. Then you take the result and say, okay, we've ruled out A because B turned out to be compatible with what we see so you can start from a biased point of view. It's what you do next that makes it science.

Dave Godlove:

You're allowed to be wrong. And not only are you allowed to be wrong, but you're expected to be wrong about 90% of the time.

Alan Sill:

And you're going to be whether you like it or not.

Dave Godlove:

You get to the point where you're just like, I'm wrong, I'm always wrong. Everything I know is wrong. And then that's when you can start to make progress.

Alan Sill:

But you can do a test and you can say if it's one way, it'll come out this way. If it's this other idea, it'll come up that way. Then you're doing science. And kids do science just naturally. All kids are natural scientists, right? We then drum it out of them. I have a theory about why this is, by the way, wrong explanations.

Rose Stein:

What's the answers Allen? Are you talking about Santa?

Alan Sill:

Yeah, partly, since kids go to school and they get told something. And, if it's a wrong explanation, why do airplanes fly? They get the wrong explanation for that. They will think, I don't understand that. So either I'm dumb or I don't like science and I don't like feeling dumb. So I don't like science. My personal campaign for this is to try to get correct explanations out there as much as possible. Because those actually make sense and that keeps the kids interested. Why do airplanes fly? They push air down. Next question. Same reason the helicopters fly, they push air down. And Newton's law applies if you have enough momentum transferring in the downward direction, that makes a force in the upper direction that's how airplanes fly. So, you get correct explanations in front of people and they won't frustrate them. So I think the way to keep people interested in science is to educate the "bejesus" out of the science teachers and make sure they're giving correct explanations.

Zane Hamilton:

So, I have a question, Alan. It goes back to something you said earlier and it ties back to what you're talking about now. Do you think the internet has helped or hurt that? Because teachers, obviously, they're not always right, they don't necessarily know the right answer. You have an amazing tool to go get information from, which is the internet. You don't always know what you're getting. So has it helped or has it hurt?

Alan Sill:

Well, yes. I hadn't thought of putting the two pieces of this argument together. I actually raised the question about whether it's been good or not. But let's just try to connect my last answer. If the internet is going to provide wrong explanations, then no, it's not good. That covers a lot of ground, right?

Zane Hamilton:

It does cover a lot of ground.

Rose Stein:

Well, there's no such thing as scientific fact. Apparently, we have discovered it over this time. So what are you supposed to do?

Alan Sill:

It's true.

John Hanks:

I think that the key is to be open to new data, and that's something that humans just are hardwired not to do.

Rose Stein:

I like that.

John Hanks:

You know, we're not open to new data. There are new views.

Rose Stein:

Hmm. So it's like a yes and not just a yes or no. It's okay. Yeah. And there's going to be more.

John Hanks:

Yeah, it's a yes based on what I know today, and we'll see what happens tomorrow going on the best information we have available.

Alan Sill:

Let's go back to the why though. Maybe, we could get a couple ideas. Brian and Forrest haven't chimed in a while. On the why. Why is HPC the best way to make society better?

How Does HPC Improve Society [41:28]

Zane Hamilton:

Brian came off mute. Brian goes first. Brian's first.

Brian Phan:

Let's see, I guess in a previous role I worked for a genetic sequencing company. Our products were genomic panels where you could screen for specific genes. One panel in particular was screening for breast cancer genes. These would typically be ordered by patient's oncologists. And, from these results, they would empower the patient to consider medical decisions to either mitigate the risk of breast cancer. Whether that's through mastectomy or other types of treatment. And from that, it could really change the course of the rest of their lives. And I felt like by building the HPC system to support that type of effort, I felt like I was making a meaningful change in this world.

Zane Hamilton:

Thank you, Brian.

Rose Stein:

Yeah, thank you. Thank you for sharing that.

Forrest Burt:

Similarly, having been on the academic side of HPC I've seen just how broad the different use cases that HPC supports in total. HPC is, on certain levels, most everything that you interact with on a daily basis. We've touched on car safety simulations, but it's also electronic design automation for chips and things like that. One example that comes to my mind that I enjoy working with in this space is the weather modeling that certain places have to do in order to determine their renewable power output that they can expect at a certain point. So, Alan, to return to your question. Why is HPC the best way to have an effect on the world? Like I said, just there's so many unexpected things that are HPC. Like I said, public utilities have to run HPC simulations to determine what the weather is going to be like so they can see how much wind power they're going to have available, or how much hydroelectric power they're going to have available.

One use case that I was particularly fond of that I got to work on, which I felt like was meaningful also in that sphere was this non-profit group that I got up and running on one of our clusters back in the day. This was very hands-on and we were setting up custom nodes for them, in-putting, ram, stuff like that. This was some of my fun early experiences in the data center. But they were doing research on renewable energy versus birds specifically in the wind sphere. So they were doing all this research into how can we develop models that determine from imagery when birds are about to potentially hit wind turbines and idle them, so that we don't have these problems with bird migratory patterns and our wind turbine farms and stuff like that.

They were also doing massive data analysis of migratory patterns all over North America. Things like that. So, ultimately just in the end, especially like I said, getting to be an academic HPC and just seeing how this group is doing this thing with weather. This group is doing this thing with some type of engineering. Just seeing how ultimately in the end, so much of what you interact with on a daily basis from flipping on the switch, turn on the lights to the hard drive, all that stuff is ultimately in some level HPC you find eventually that so much of just the basic stuff that we do in our global civilization today requires such a level of extreme computing for us to be able to figure it out. It really becomes quite interesting to see just how far reaching HPC is, and like I said, unexpected use cases, renewable energy, nonprofits doing research that you wouldn't normally expect to be in that heavy computational sphere. It's incredible what HPC actually touches in the end.

Zane Hamilton:

Thank you Forrest. And I know we do have a question from Steve Moody. What are some real world examples of purpose driven computing initiatives, if any, that have successfully integrated personal motivations and values? Give you a second to think about it. I'm going to pick one of you that looks like you might have an answer. Dave, I gave you zero time.

Real World Examples Of Personal Motivations In Computing Initiatives [46:00]

Dave Godlove:

That's a difficult question. So integrating personal motivations and values into purpose-driven computing initiatives. Well, I mean, I don't really know if this is really what Steve's asking if this is, but once again, going back to science I would say that a lot of times people's personal motivations are deeply tied to the types of science that they're doing and therefore the type of computing that they're doing because they're using computing to analyze their data and doing stuff like that. But in those cases, the personal motivations align with people's career trajectory. Because a lot of times scientists honestly have a personal motivation to be great scientists. They want their scientific discoveries to be disseminated to every corner of the world and they want it to become the basis. A lot of scientists are very personally driven to achieve and probably not just scientists, but probably in lots of different disciplines.

People have a personal motivation to really achieve things and to really be known. It can be both a pro and a con because sometimes people take shortcuts and do things that they shouldn't because of that. I would say that the area of science that I come from, for instance, I don't know if this is the way that it is in a lot of other disciplines, but the area of science that I come from there was that almost people were actually frowned on if they wanted to make money. Making money was really a knee jerk negative reaction to it. And the reason is because people felt if you were motivated by making money, then you weren't doing the kind of caliber and quality of science that you should be doing. So everything people were really excelling in their careers and they were doing so because of personal intrinsic, not extrinsic motivation they were getting. I don't really know if that answers the question. It's just kind of my hair brain thoughts on it, but I'll kick it over to somebody else who might have a better answer.

Zane Hamilton:

Good. John, what do you think of this question?

John Hanks:

It's really a hard question to answer for me specifically because I got into this after realizing that I really suck at wet lab work and that I'm never going to be smart enough to come up with my own creative ideas to make the world a better place. And my sweet spot is helping other people carry their ideas across the finish line. So for me, there is no big project that involved any of my personal motivations, but I've had a chance to put my dirty fingerprints on a lot of stuff that has been well received and helped a lot of people by carrying my little part of it. The people that I've helped though, they have done exactly this purpose driven computing to carry their personal projects and their personal goals forward. But, I wouldn't want to speak for them.

Zane Hamilton:

Thank you, John. Alan?

Alan Sill:

Yeah, so I feel like I can combine the previous answer a bit. I did set out to excel in the discovery aspects of the science I was after. And I had some personally very satisfying experiences from that. But I'll try to give an example that came up along the way. It's something we've talked about in past webinars. I'll just give a precise example. In the process of trying to do the computing, to look at the collisions at Fermilab and in the pursuit of the top Cork and other particle physics, we ended up needing a lot of computing. And this led to a project that initially I was the only one in charge of, which is to teach people how to duplicate the clusters we had built at Fermilab for using Linux, in anger, if you will on purpose to build dense clusters of computing.

And these got duplicated all over the world. And so it became a project to try to coordinate these initially just within our experiment and within our field. And then I joined projects that led, along with the effort of many other people, to the creation of grid computing. The idea of using cluster computing for multiple types of science. And so these computing initiatives eventually tried to pick up. The second part of the question integrating personal motivations and values. One of my motivations was to get people to work together. And it still drives a lot of my own work to try to bridge barriers, cross borders, work in an international and cooperative way, wherever possible. I still believe in the positive aspects of science, and I just try to be a little more aware of the additional negatives that we discussed.

So, the creation of large scale distributed computing is something I feel I've participated in. I don't get the personal satisfaction that I got out of the science work, but I feel like I've done that. And now I can go on trying to use my personal values, which are communication and cooperation in international ways through distributed computing. And, you folks who are helping other people through HPC, the implication is that it's something that people need help with, right? That they can't just snap their fingers and open a browser tab and poof there it is. It's something that requires talent and expertise. So, there's a sense of satisfaction that comes from that as well.

Rose Stein:

I just want to say that is a good idea though, Alan. I like where you're going with that. You open a browser, right?

Alan Sill:

Yeah. That's right.

Zane Hamilton:

Sounds interesting.

Alan Sill:

Where is the downside, right? You don't want it being used to take little old ladies' mortgage money, right?

Zane Hamilton:

No, absolutely not. We do have another question.

Modeling And Simulation Application Use In Experimentation [53:10]

Rose Stein:

David, can you touch on the savings in material fabrication, animal testing, destructive testing, traditional experimentation that modeling and simulation enables? Ooh, Forrest?

Forrest Burt:

I don't, not to take this a whole different sphere of use cases than we've discussed so far but, along the lines one interesting example of this that I ran into was in the national lab space. While working in the Idaho National Lab sphere, in terms of national labs, they work a lot on nuclear fuels and waste management and stuff like that. So developing safer fuels and developing more efficient nuclear fuels that type of thing. I worked at one point with a PhD student who was doing a lot of simulation. I believe it was the cladding for fuel rods or something similar to that. And the benefits to material science in not having to build out all of these hypothetical materials that they were wanting to look at.

So like exotic compounds, ceramics, metals, these, building materials, that we would expect to be involved in the harsh environment of a nuclear reactor. Novel materials that would take significant laboratory time and engineering to be able to build a process to, what's the word I'm looking for? A chemical engineer, basically to get to actually industrial scale production. So they were able to, or she was able to take a lot of these hypothetical compounds, like I said, ceramics, building materials, stuff like that, and simulate them very effectively on high performance computing resources without ever having to have all the costs and everything associated with actually, like I said, going through all of the prototyping and industrial chemical engineering and material science engineering and stuff it would take to actually bring some of these materials into reality. And so they're able to just take the most promising candidates and then test those with actual physically produced samples as opposed to having to do that for hundreds or thousands of different samples. So yeah, it's very, very interesting stuff there and a lot of different spheres.

Rose Stein:

Yeah. Thank you, Forrest. I know Dave, you did a cool demo. A crash testing demo for us at one point. Do you want to speak to that?

Dave Godlove:

Yeah, I mean, so that was OpenRadioss, which I did not do the work for. I have to give a shout out to Yoshi. So Yoshaki Senda is the person who was actually responsible for containerizing OpenRadioss and then putting a blog post out there showing how to do it and everything. But yeah, I mean that program is a material modeling program in which you can deform materials and break them and do things like that. And its most widely used application is within the automotive industry. As far as, and I was alluding to this earlier, but its most widely used application is to take and build models of existing automobiles with as many different materials as possible.

So that you have the engine block and you have the material that the engine block is made out of. You have the glass windshield. You have the metal hood. The plastic bumper and to as minute a scale as you can all these different materials. And then to take that and to run it into a wall and see where the stresses go and to make sure that you are taking the stress and absorbing it in the front of the car and channeling it around the cabin and doing things like that to make cars safer. And so you can appreciate how expensive it is to build a car and put some test dummies in it and to go run it into a brick wall.

And you get to do that once, versus if you can actually model a car and put it all together in silicon and then do that over and over and over again. I will say however, that's a really, really great application of modeling and simulation because we understand so well what's going on, right? We understand so well what the materials are made of and what their properties are and how they react under forces and under pressure and stuff like that. A lot of science deals with systems that we don't understand well, right? And so when you get into systems that you don't understand well and you're doing the science to try to better understand the system that's where modeling and simulation can be helpful as a tool to help you better understand the situation.

But it's not something that you can rely on in lieu of doing the actual physical research. And I bring that up because one of the things that David is asking about is animal research, for instance. And I used to be in animal research and that's one of the questions I would get sometimes. Couldn't we just simulate the studies that you're doing instead of doing them on real animals? And wouldn't that be a lot better and a lot nicer? And the answer was well we could. I was doing research on the brain. We could simulate the brain if we understood how it worked. Since we don't yet. We can't really draw meaningful conclusions from the simulations we do because we don't know if they're based in reality.

So while we can take the data that we have and feed those into simulations, so there's a feedback loop where you collect real data from the real world, create a simulation, feed it through and then make predictions based on what that simulation shows and then go back to the real world with those predictions and test them in the real world. And that feedback loop is helpful for making progress but to just remove the real world part of that and just do it all in a simulation when you're talking about systems that you don't really fully understand yet, that's not really as helpful as it could be otherwise.

Alan Sill:

Sure. It's partly a point of efficiency. I put in the comments on YouTube that the best talk I ever saw on this area was from the head of the research computing for Proctor and Gamble. A very old company. More than a hundred years old but very modern computing. When asked why they spent so much money on computing, he pointed out the tremendous savings with that feedback loop in mind. You can engineer the example of a jug of liquid laundry soap, right? They sell a lot of those jugs. Well, they want to make sure if you drop them it doesn't break and spill laundry soap everywhere. So you can make your molds and make jugs and drop them. You can drop a hell of a lot more laundry jugs in simulation for a lot less money than you can in the real world. And so this feedback loop that you mentioned is crucial but HPC makes it efficient.

John Hanks:

There's a, I guess we started out, I was thinking that this in terms of what product of HPC is going to make the world a better place. But this question drives home the fact that there are lives and suffering to be saved on both sides of that product and specifically in pharma and in drug development and medicine. Well, I'll sidetrack here for a second. One of my favorite comments from the beginning of a talk that was given was there's good news for all the mice in the audience. We've cured cancer.

John Hanks:

Pointing out that for mice there's no reason for a mouse to ever die cancer again. We've solved that but that doesn't help humans at all. And the idea is it could. If you could cure cancer, would you do it? Yes. Well, if you had to kill a billion people in the clinical trials, would you cure cancer? Well, no, probably not. And doing molecular dynamic simulations and doing some things with stem cells and Petri dishes there are a lot of techniques that all can come together and reduce a lot of that upfront suffering. Weed out all the drugs that aren't going to work out early and focus the actual suffering that has to take place on just things that are likely to work. And I think that's probably one of the most important aspects of computing that is done just to reduce the amount of suffering going into achieving a goal.

Rose Stein:

Yeah, good point.

Alan Sill:

That puts a bow on all this. People ask what's HPC good for? Well, it's a very long list. And, the idea of making advances more efficiently while, again, being aware of the potential negative outcomes is just a tool we use. But it does make this whole process move a lot faster, at least, hopefully.

Rose Stein:

Yeah. Thanks Alan. All right. We are just about at time, but Zane, we have a nice solid minute for you to close us out with your why.

Zane’s Why [1:03:12]

Zane Hamilton:

My why. So, my why is a little different because I was not a research scientist but I spent a lot of time in the airline industry early in my career. And once I started learning more about HPC, I realized how much stuff could have really changed in the industry if they would've looked at doing things via HPC instead of trying to do things the way that they were. If you look at what happened around December with the airlines and the tough time rescheduling and the outages, a lot of what's done there is not done on HPC from a scheduling perspective. It's usually done on one or two machines. And if that would've been looked at from a much larger scale a lot of those models take a month to run to get an optimized version of that from a fuel and from a personnel perspective, if you look at it from HPC, you could have done it in minutes or hours and been just as efficient as that one machine running for a month.

And then, I mean, if you let that run for a month, what would that have given you? And then being able to change as you're having issues and run a model that you could have a better solution to a much bigger problem in a matter of minutes. That became very important to me. And then when I started spending time with people like Dr. Godlove here and talking about how it helps in the medical world and the things that you can do and really help there, I think everybody is touched by that in some way. I know I am. So I think it's interesting and very helpful to the world to be able to do that. So guys like John, I appreciate it. and I appreciate everybody's feedback on this. It's been a lot of fun. I love hearing these stories. It's great to talk to you guys.

Rose Stein:

Well, yes, thank you, Zane and everybody for showing up and all of your amazing work and all of the work that you enable with your amazing work. Thank you for being here. Thank you for watching. Make sure that you like and subscribe and leave us a comment throughout the week. We definitely are going through and answering questions and reaching back out to people. So if anything is coming up for you, doesn't matter if you watch live or not, go ahead and leave us your questions there and your comments. We'd love to hear them. And then next week, I just have a little teaser. Make sure that you come back live next week and watch us. We have a big reveal. We are releasing a new product that is just going to blow your mind. So make sure that you're there. Thank you so much.

Zane Hamilton:

Thanks everyone.