 So welcome. I think we have an exciting panel today at Purdue. We'd like to do things at scale So today we're going to talk our panelists going to discuss Compute and data at scale. How can we unleash the power of these technology? I'll start by introducing briefly our panelists and then we'll jump in so we have our Distinguished visitor from Sandia National Labs Dr. Jackie Chen. She's a distinguished member of the technical staff at Sandia member of the National Academy of Engineering she leads a group at Sandia developing a direct numerical simulations of fluid and combustion and we have the Pleasure to listen to her talk yesterday. And if you missed it, it's available online So please go and check it out She's also a fellow of the Combustion Institute and the APS The next panelist is Carlos Kahlo. He's an assistant professor of mechanical engineering here at Purdue His work is on acoustics and turbulent flow and he's the founder of a startup company called Hysonic Next in the panel is Arisu Ardenaki. She's an associate professor of mechanical engineering here at Purdue She's interesting complex fluids and multi face flow She has won several awards in Society of Woman and Engineering She scanned a career award from the National Science Foundation and the presidential early career award that President Obama Gave her John Poggy star next panelist. He's an associate professor of aeronautics and astronautics here at Purdue He's interested in experimental computational and theoretical Fluid dynamics runs the whole range. He's a fellow of ASME and the American Institute for Aero and Astronautics and Last but not least is Professor Charlie Bowman. He's a Showwater professor of electrical and computer engineering and biomedical engineering here at Purdue. His interest is in computational imaging and sensing and his his group developed the first commercial model-based reconstruction system for medical applications for tomography and he's a fellow of several society so without further ado, let's let's get started on on the topic and I'm going to sit and what we're going to talk about these The opportunities and the challenges that present because of the convergence of Cyber infrastructure including high performance computing systems Communications ability to do cloud computing data repositories together with software that can make use of these infrastructure and And what we want to see is talk about the future and the opportunities and the and the challenges that these presents So I want to organize these in in a few things and and I'm looking forward to thought-provoking comments and maybe disagreements and Well and and the back-and-forth between the panels so so let's get started the three themes are roughly speaking Like to discuss the technology that we have today The what opportunities that enables the ability to complete that scale What are the challenges to democrat democratize this technology how we can put these tools to the hand? of not-academicians and a few research groups but to a large group of folks who can make use of them to benefit society and What are the challenges in education so how do use this technology to? Develop the next generation of scientists the next generation of engineers So so let me start briefly by saying in The fastest super computer in the world in 1997 was ASCII of red at Sandia National Labs. I'm sure Jackie used that computer And that was the first computer to break a teraflop. Okay, that's 10 to the 12 floating points operations per second That was in 97 Today we have teraflop power on our desktops We we have millions of those computers Installed in the world We have the next level up in terms of simulations a thousand times more powerful Took about 10 years to develop that was the roadrunner computer at Los Alamos and that was in 2008 Now we have about 1000 teraflop computers in the world and we're going to the excess scale So as the leadership computing goes to the excess scale, we have a thousand or so teraflop systems and we have millions of teraflop computers and My first thing what I'd like to hear from the panel is what opportunities That's these enable in terms of compute and data. Okay, so maybe we can start with Jackie So my area of expertise is in combustion in turbulent rafting flows and I think what the current About 200 petaflop machines. I think some it is the world's fastest Machine on the top 500 as of a couple months ago. It's allowing us to Enlarge the range of dynamic range of turbulence scales that we can simulate To and include some degree of complexity in terms of multi physics. So not only can we start to simulate Detailed reaction kinetics that are relevant to the practical fuel that we use in In our cars and automotive engines and in our power plants that generate electricity But also I think having that kind of capability lets us compute with extremely high fidelity Scenarios or configurations that are relevant to both experimental laboratory flames as well as to to starting to become relevant to industry so configurations that represent processes combustion processes and IC engines or in in gas turbines and So this is a really great opportunity to combine experimentation and high-performance computing and simulation and Design these numerical and physical experiments from the ground up as a group collectively so that we can clean more physical insights and Provide data benchmarks both computational and experimental that industry and students and and Folks from other Institutions can use those data sets and benchmarks for validating their models or for their own purposes So things like developing portals and gateways That are accessible by the broader community and Having access to the data as well as software tools to manipulate the data would be something that we're kind of writing starting to do now So so let me bring Charlie for a second. Then can we talk about? Access to Distributed computing not necessarily the leadership type computing that Jackie was mentioning but having lots of powerful Systems to bear on the world that you can have access to maybe in the in the medical field and have an impact yet Sure. No, I and I think Purdue is actually a good example of that because I mean at the risk of sort of advertising some of the things we've done We've developed this the cluster computing system, which yes, which is very much democratized computing on campus by allowing Individual faculty members to buy into a kind of an integrated cluster and really reduce their costs So cost of computing is really crucial in terms of getting it out to users and it allows people They're non experts to get derive a lot of benefit and we've done that with both CPU based systems and GPU systems and And health care is a good example of where high performance computing has played a huge role And we'll continue to play a huge role and where cost is an important factor So it you know unlimited cost is not practical on health care. So so but I think the community in health care has Technologies like CT and mr really bring together physicists and people who develop algorithms and computer scientists and biologists and and Medical experts to solve problems and more and more they're realizing the computation is a really key piece of that Okay, and what they're and what you can see across the industry a key Direction is machine learning and what they call AI which is playing a huge role And it's going to have huge computational demands and require different kinds of computing platforms So it's definitely the case that because I work closely with various health care commercial companies and there's a huge push to integrate AI into Imaging applications in health care and the last thing I want to leave you with is that inverse problems as opposed to forward modeling so What professor chen's work did really was beautiful work in in simulation of combustion and Ultimately engineering problems are to design systems that solve problems and inevitably you need to solve inverse problems Which is a particular interest of mine, of course So so that's going to be I think another big push how you can solve it put those forward models into sort of larger computing problems systems that solve the Inverse problems required for sensing and design So I guess more on the controversial side When I when I you mentioned the explosion of computing power and computing capabilities, and that is that is that is a hardware And a hardware technology that is hard to keep up with right and so it's overwhelming from a computer physicist side to do that And as Dr. Chen mentioned yesterday, it's hard to keep up You need to have a team of computer scientists it becomes a different job at some point and and I think what is more humbling is To think of a super computer perhaps like the human brain We all know or they say that we only use a certain percent of it I have the feeling that our codes Do the same right so there's been a recent I cannot recall the name of this ETH group who? Re-engineered completely their codes so that they could use every single cycle of the CPU because otherwise our standard legacy codes Only used maybe 20% and for the 80 for the rest the CPUs were dormant And so that it is these are very exciting times But they're also terrifying from from our perspective. We're not from a computer physicist perspective So there's a lot of power out there There's a lot of potential, but I don't think we're tapping into it And it's becoming almost distracting from our work, right because it's very hard to keep up So in a way, it's a word of caution that I might raise here Yeah, so in some ways I mean to again to take a controversial standpoint Supercomputers are getting worse. So for in terms of traditional computational fluid dynamics techniques We are seeing a drop-off in per processing in the element speed So very large computers are great for us in the sense of capturing a separation of spatial scales You get more and more turbulent scales in a computer simulation, but we have a lot of problems that are stiff in time And the only way to crack those problems was the faster and faster programming processing element And modern computers are getting a bit slower in the last few years. We've seen slower per core speed In order to reduce the power consumption So we need to like Carlos said rethink from scratch the approaches We use to make use of these this hardware because we're not going to see a miracle going back to faster and faster core speeds Just because of this cost in terms of power. Yeah I wanted to add as we move from Megaflop computations from In 1970s all the way to now talking about exoskeleton pitting what also matters is talking about What are the most important and toughest problems in the world that you want to tackle with we heard about Healthcare we heard about combustion and how that's related to emission and other challenges We are dealing with but there are two areas that I wanted to just Touch on which exoskeleton pitting can impact on one would be on Climate forecasting prediction of hurricanes air pollution Ocean pollution and the current models have resolutions on the order of 50 kilometers 100 kilometers where they would develop basically earth system models which includes Chemistry physics chemical evolution and everything together, but we now know clouds low clouds convection processes Ocean eddies all of those would also Contribute to those on climate models, so it becomes important to use these exoskeleton Computing powers to go through resolutions down to even one kilometers or even blow to include some of those effects One other area that I wanted to mention is on biology of where Resolving different scales becomes very important going from atoms to DNA to South Oregon to organisms and understanding these Metabolic processes these cellular processes and these complex interactions become important. It has been only a few years ago where million atom Computations in biology being possible even though those being possible material science for other areas decades on before and that's because we still need mathematical tools to be able to include these complex processes and just to add a question on that regard is that how What we need is basically a rigorous course on graining top-scale these processes How are being that regard and what are the challenges that we are facing in developing those course brain Yeah, so it seems clear That that we need teams of people right and then the national labs are particularly good at these where they can put together teams of Computer, thank you computer scientists domain experts Your experiments for for validation and that's really what would allow us to solve You know these problems that we're discussing in my field is material science and we have the same Multiscale problems the same time-scaled problems where you cannot easily Paralyze time which is what John was mentioning, but there's new algorithms coming up where You can use statistics methods to achieve that parallel replica Approaches and whatnot. So it's really a combination of hardware new algorithms new ways of thinking about old problems, right? That's going to really Help us make use of these leadership type Computers but Moving on maybe to the to the second theme in terms of democratizing access to these tools what I So as we have these teams, right? We have experts in computer, you know computer scientists domain experts I'd like to think a little bit or discuss a little bit about the end users So I think if I have to do a criticism a self criticism of our field We end up often being the end users of the products and the leadership type computing serves a relatively small group Of people around the world and we have thousands of peda-scaled systems and and millions of their Terascale systems that could be better utilized We all have Smart phones and you don't need a manual to use that right and it's very sophisticated technology You don't lose files. You don't organize. You don't have to create directories by hand the way our students do when they organize HPC Systems so there's a whole Set of technology in the commercial set sector even micro services, right? Netflix and a bunch of companies developing very sophisticated Systems that I think our field Could benefit from and you know make us better developers better users and also be able to transfer our technology to end users that maybe Maybe engineers, maybe doctors who can benefit from these tools without knowing the Inner workings of the two knowing the physics and the application the same way one can drive a car without knowing the inner workings of the engine so Thoughts on on that democratization Completely agree with what you said But the data that And there are our machine learning tools and AI tools that are allowing To develop surrogate models or digital twins models That are far simpler and less complex and smaller and faster to run once they're trained on a few selective experimental, you know data sets or for our hero simulations and so using those on on More ubiquitous machines like tear scale or pet a scale machines in a few years would would open up access to the data information and allow Much larger parametric sweeps and optimization problems that for example industry would care about if they're trying to design a Product or or engine in my field. I think having those tools and then having workflows, I mean part of the problem with with sharing data has been in the past that composing dynamic workflows and And and bringing the community on board only works if if the software is not buggy I mean most people are willing to give it a try, but if they're stuck with things not working as they're supposed to They're only do it once or twice and then they move on and give up and so having hardened Software tools and they're like you said there's a lot to be learned from the commercial sectors, right? That are already doing that and data as a service and all this kind of thing and applying it to the science and Fields is something that we need to to kind of get a better handle on Let me make a point that we're going to have questions from the audience specially students So start thinking about those questions would like to open the door in a minute Which is deep learning everybody knows about and it's up. It's been amazing to me Very surprising that for a lot of problems where I thought you needed highly Accurate physical models that these machine learning methods particularly deep learning Can often replicate? performance that's a quality fidelity that's Comparable and when you consider how much faster it is then you can incorporate more effects And you can get actually better results so in other words you can use them You can basically train the data the machine learning models with a higher fidelity model So that in practice it gives you high it may give you higher fidelity than a physics-based model, so I Mean that's sort of all shaking out and maybe it won't quite work out that way But I think that we need to be Sort of cognizant of that and how it can be leveraged Maybe where we have as you suggested, you know very high fidelity models that run on a few Computers and maybe that data is distributed widely for use in training machine learning models I'm not exactly sure, but it just seems like this is kind of a game-changing technology Right and maybe Include is incorporating more physics-based Informed machine learning yeah, and there's a constraint it and also take care of you know There's new met math methods to help identify rare events or right extrema you know data out in the tails of Absolutely, all that still needs to be done. I think people are still grappling with how to incorporate physics And there's likely to be a lot of innovation Sometimes I think about is to be we rely very heavily on machine learning techniques and treat Physical processes as black box and come up with surrogate model and not include any physics-based model Are we losing our fundamental understanding of physics of more complicated processes? So I think it becomes very important to look at both aspects together as we evolve newer and newer methods To add to that you know machine learning basically Interpolates between physical model cases that you have computed So if you step outside the box you've calibrated your machine learning model your your model will not be correct And you have no idea what the error bounds are on a prediction that's outside the box It could be anything so machine learning in that sense is extremely dangerous And if you'd use for that for something critical like medical applications, I'm a little bit worried about Like diagnosis by machine learning Society's done many extremely dangerous Yeah, I mean I totally agree with what you're saying and that's why I entitled this talk fair the deep but But it's but you can't on the other hand ignore The other comment I wanted to say is I have a sign in my office that says deep learning leads to shallow thinking so but that You know Infra physical science has for the last you know for a long time been very Experimentally oriented in much of its endeavors and but information science has been much more kind of theoretical analytical And it's moving away from that so it's more experimental information-based science. So And that's going to be something we're gonna have to get when we're more comfortable with I think if the data was out there you could train these models with data that's available Machine learning can actually tell you about Outliers and about the model like Jackie was saying that for whatever reason is giving you the wrong values and It seems to me that as a research community we spend sometimes, you know millions of dollars and All that data ends up in a PDF in a paper. That's not discoverable It's it's not that you cannot query it and it really contributes very little To the knowledge and now we have the infrastructure to actually do this and I think that would allow us to train the the machine learning models better and also Assess, you know uncertainties and assess where we're extrapolating and we're we're you know in dangerous territory enthusiastic about is Using it to find surprising things in large data set So one of my problems is I have large data sets I had to recently did a calculation with a restart file for a computation with 3.2 terabytes every time step I regenerated that and I couldn't possibly save all the data So I have to take a guess before the computation what will be interesting and save that the machine learn now Learning algorithm could tell me well, maybe you should be saving this and that's that's I think is extremely I think computational steering is a really nice use of machine learning and things like anomaly detection right using taking advantage of Information for example in the higher moments beyond the mean and standard deviation might allow you to identify Rare events or you know something that is an anomaly and that might steer increased IO or in in our combustion field Maybe it tells me I need to inject more fuel to get things to light up faster or something right Related to the democratization of high-performance computing There's there's clearly a spectrum of problems you can run things that are called embarrassingly parallel and then the end users could be And which includes part of machine learning and then say Monte Carlo simulations or data mining and then all the way to something very technically specific that might have a lower impact such as Aegon value solvers solve them massively parallel architectures or even DNS and LES that might seem obscure to most of the community out there and so so when it comes to democratization, I think there's a good chance of Making HPC resources more accessible and democratic if you think of cloud computing where there's that level of abstraction Where the end users which could be companies which could be somebody on an iPhone in the future They don't have to worry about the hardware It's it's there's a layer of abstraction, but then that's one side of the spectrum on the other end of the spectrum There's the computational physicists that have big data fever and they want to dig into the hardware They want to optimize their code to the last bit And so I think it is important to keep in mind that there's a spectrum Right it goes along the software stack from the application on down to the hardware. Yes, and so I think HPC can be Will it ever be more democratic more accessible it depends you need to work your way from the spectrum from one end to the other of the spectrum, but I think we're still I Don't think that the end users are still There should be no reason why the iPhone couldn't make use of Cloud computing nowadays, but I don't think we're there yet So I have to put a plug here for nano hub by the way, which Have about 500 simulation tools that are all web-enabled That that run on HPC resources, but you can run a simulation with a few clicks from your iphone This was not planned Out of an idea that mark lansstrom had and a request from a colleague who wanted to run a Simulation and here the student put it up and make it available and then a Combination between marks idea and some folks at the National Science Foundation What the guts to see you know, thank big and say this could be turning to something. That's a global cyber infrastructure and now it serves millions of visitors every year and Tens of thousands of simulation users including your classroom usage of really sophisticated tools but that are Simplified in terms of the interaction and abstracting away all the computer science the algorithms You need to know the physics Let's see we have a question from the from the audience any any questions from faculty or students team With democratizing Technology and all throughout history we're from three computers in the world take up a whole room No one could see exactly how that would change a lot of people's lives as we went along and that trend has continued So it's inherently hard to predict, but I'd like to kind of put it to all of your expertise and imaginations. How do you think? democratize pet a scale and maybe even some they exit scale computing to everyone Might open up things that are hard to imagine now, but could have a really deep impact on civilization If you look at how cloud computing is done and I did some research on that Yesterday because I was not involved in cloud computing I work with my own little world as a computational physicist with my cluster and Classic clusters and architectures and the first thing I thought without reading anything is like oh my god Heterogeneous architectures a computation is run a little bit in Africa a little bit in Asia a little bit in Japan and then data is collected together I don't want to deal with that, but obviously there's a there's a nice layer of Abstraction you don't see it what what is presented to you as a virtual machine So you don't even know what's going on. It's Amazon's problem or somebody else's problem So can access scale reach that point where the user doesn't know it doesn't want to know and should not know that Probably it's already they already do it with cloud computing and I would have said that was impossible to do but so future is bright, I guess It's great to do the democratization and have everyone access and Unleash the power of existing computing for many different aspects that we can't imagine now But again playing it that will add with it here. What if also people who don't have good intentions Get their access that and use it for things that are not good for humanity So that's also the other side of it to think about Security and authentication will become even more important Future from a lab standpoint, I was checking the you know the amount of compute power in the commercial sector and You know in a cloud Service providers is increasing right and I don't have a crystal ball, but you would imagine that the times in which National labs completely dominated HPC might be Passing and there may be serious competitors in terms of HPC in the commercial sector. And so do you imagine? You know, this is true also for companies What would be required for an organization like a national lab to run maybe the non Not super sensitive computing outside in cloud resources and if anyone has some insight from a commercial standpoint from a company standpoint What's what what are the barriers to? Doing some of this computing in the cloud Algorithms are more amenable to cloud computing where you have looser Connections between processes right where they're embarrassingly parallel and there are other problems like partial starving Differential equations that require much tighter coupling that probably won't happen very well Necessarily in the cloud unless they have very fast interconnect Yeah, and I said I should say one of the things that has to happen is You know even as computing gets faster and faster a thousand fold every couple of years The networks also have to increase in speed to accommodate the higher throughput and bandwidth Right so right now things like the energy science network that DOE puts up is maybe a couple hundred gigabits per second in terms of data transmission within a Large part of the US into Europe But I think if as more people Hop on to this And do computing in the cloud or streaming data You're gonna have traffic that network that's gonna be the bottleneck and not the computing itself And that's another so that's another infrastructure issue There's an interesting challenge and for example in health care in like It's sort of moving in that direction, I think that you'll start to see maybe Not you know sort of local servers in hospitals and do that sort of things that they get better utilization the death advantage of The of having embedded computing high performance embedded computing with each You know device is that that thing is sitting idle most of the time, right? Right, so it's very inefficient, but but the advantage is that you don't have to move so much data around and you know one example of It that you know, it's had big impact is on our cell phones with speech recognition, sorry But anyway, the you know, that's an example sort of a combination of local and server cloud computing that's enabled a very impactful Application but you know the challenge there is the balanced amount of communication and also the security issues Moving people necessarily want the secure data distributed all over the world so And yeah, so it's gonna be a real challenge often cost right now I think it's high for like for instance, we've looked at the cloud services on Amazon for secure computing. It's very expensive Yes methods for multi resolution visualization for example where Reconstruction of data right so that you can ship small packets fast and then right absolutely and only the critical information Right the Colossi methods which identify critical information and just transfer right for even parameters of it That's probably interesting because there you need those type of techniques also in leadership that computing because you can't store That you cannot store it right and so maybe there's an opportunity there where you can use similar techniques on both ends Where you do some local processing to ship the data This communication problem actually brings in also the question of reliability Systems so if you're looking at millions of processors and your chance of losing a processor is one in a million Is it probably going to lose some processes so being able to proceed with a calculation having lost some of it And maybe do some kind of error correction to keep up might might also be a big breakthrough would would also help with the communication I think along those lines developing math algorithms that are resilient to Failure right or iterative methods are yeah can pick up can keep progressing when there's errors or asynchronous Yep So some combination of local computations of speculatively and in a correction from the global network of computation Yeah, we have another question from the audience another student Always a student so I was fascinated by the the two bookends that have come up in the discussion one on the side of accessibility of You know to and pervasive computing and accessibility and on the other side you're talking of leadership You know computational tools and I couldn't but help think of you know the path that most disruptive technologies have followed and One that I've been reading about recently is gene drives for example One interesting thing that you see is the development of the what you would call leadership kind of systems Really expensive, you know just hero, you know heroic computations kind of stuff is always sold based on potential impact to You know key really important problems the gene drives are very much sold as G we might be able to create a mosquito that you know knocks up all the other mosquitoes that you know spread malaria You know so there's grand challenge problem. Everyone gets it Investment follows but then as gene drives come, you know prices come tumbling down accessibility becomes easier Everyone is able to use them to their locally defined, you know Challenges so it's it's probably very clear that as high-performance computing whether cloud-based or you know Otherwise hybrid systems that it goes down to the hands of users will define our own ways You know in different places to drive change and transformation But at the leading edge of it, you know All of you have been in situations talking about it. What are some of those grand challenge impact areas that you think? You know to say something like hey, we can get rid of malaria. You know that level of impact What what what can we say in the different different fields that you're in some examples of those kind of? impacts Maybe I'm sure Jackie Sounds like a lot Calibrate right it doesn't solve World's problems, but I think what in in our field in combustion what it will do is like I said before it allow Us to incorporate a couple simulation and experiment to provide as much detailed information In as relevant as post, you know practical engineering spaces that can inform industry People designing gas turbines and engines and and then maybe through machine learning and other types of methods provide surrogate models and better modeling that industry can then use to Run millions of calculations to help optimize their design of both the fuel as well as the conductor It's interesting. I mean I'm gonna actually push back against that hypothesis a little bit I mean we tend to want to look for the big application and and they will and they're very important But sometimes what really has impact is just making something simple more efficient like and I think you know the CPU was a huge Revolution in computing because it's sort of a commoditized computation right we took a lot of complicated What flip-flops and gates and so forth the bits like processors and we said okay We're gonna have this single generic CPU with the defined instruction set And we'll just keep making it cheaper and cheaper and cheaper and so more important people have it a similar thing happened with Machine learning one of the reasons I think it's so hot right now Is it a lot of the technologies existed for quite a while, but they just took some of it? They took a few key modules that were well understood and they made them really efficient and fast and cheap and easy to use Okay, so I kind of think with high performance computing Maybe what needs to be done is identify a few core Operations that are sort of it very commoditized and easy to use and then they could be sort of democratized Across a lot of applications and then we find out what the big wins were when people tried a lot of different things So it's hard to say. I'm not saying that that's right, but it's not necessarily wrong Using these hero machines with different types of heterogeneous processors tensor tensor processing units GPUs CPUs What we're seeing more of at these large-scale runs is combining physics runs with machine learning in situ Where you devote the tensor processing units maybe to convolutional neural networks and then you use you know PDE solves on The CPUs and maybe chemistry solves on the GPUs and so as I think you said pointed out earlier You know, we're only using a fraction of the machine to do the actual Science solve and there's all this extra real estate on the machine to do these analytics and machine learning types of things and couple them together and said to couple it together effectively you need to have a runtime or a Programming model that that kind of sees sits on top of all of it and kind of can orchestrate Holistically the data movement and which resources different computations should be computed on dynamically and and even having that kind of Software stack and runtime system. I think at the bleeding edge will trickle down to Well, this is a forecast Don't know if it will right to more common things that we're all going to be doing right trying to grapple with People do things right doing data-centric computing You know, maybe I don't like to bash MPI but But it sits at a low level and maybe it's seen its better days given the changes in hardware and and the use modes Another question Good comment on two phrases analog computing one phrase and also comment on another phrase artificial intelligence So we've talked a little bit about artificial intelligence with Charlie and maybe we can discuss a little bit neural computing neural processing units that are Analog and They're used, you know, they're on on devices today to do parts of the computing Yeah, no, it's it's really hard to say. I mean it's interesting I started my career at Lincoln lab working in an analog signal processing group that we're using surface acoustic wave devices and and sawfets and And but then what happened is that the wave kind of pendulum swung and people said oh gosh You know you want to digitize things as close to the sensor as possible and process everything digitally because speeds were so much Higher and it's so much easier the program than it is to try to design a hardware device But you know, maybe that's going to swing back Although if it does, I mean we really we need to define where very well-defined modules in my opinion That for which you would be able to plug in that analog computing and it has to be a big win because the complexity of it is high Yeah, and as far as AI goes the you know, it's interesting. We have this huge Burgeoning interest in AI and I think that's a great thing, but really most of it comes down to machine learning There is there's an alternative argument to make which is that we're just at the we're just scratching the surface of AI We now okay these really hard problems like object recognition that we thought oh We could finally crack that but AI would be so much easier, but then so we've cracked that problem But now we realize gee okay, so we can recognize objects in a room But how do we make decisions about how to do things and that like humans do and we suddenly realize once again that There's a huge number of problems in artificial intelligence that we have no idea how to do so I think That we had a big breakthrough, but I think we're just starting a journey So so maybe as a final theme We're at at a university produce a big large university. We educate a lot of students would like to reach out to more students and Certainly all our alums how what are the implications? In terms of education. What are the implications in not just in different in educating next generation of students How can we use this technology to make education better? What do we need to teach our students? I think we should Large-scale computing into the curriculum So the student who's going to school now will probably even if a non-specialist will encounter large-scale computing in the work Environment when that person gets out and so we need to develop that sort of skills To sensibly use use large-scale computing when they get out and I'd like to add you know again It's devil's avid kid most large-scale computing is used very badly So if you look at the queue on a large-scale supercomputer, there are a lot of small jobs running it at one time That's a total waste of money and electricity. They should be running very large problems So using learning to use these kind of resources sensibly should be part of the curriculum And the point I would make it again to be a bit controversial remember a colleague of mine saying I'm not going to mention who who was Say well, we were trying to teach these students Computation physics and We open an X term that they they were all surprised and they didn't know what to do with an X term And maybe we shouldn't open an X term to begin with right It's maybe we need to think about using better tools to You know and educate students with more modern tools and not the way we used to do Computation science in the 70s right creating directories by hand and vi and all that But I love that that's how I do it But I don't think our our students should do it the same way I did but I think there's a preliminary step Sometimes we as instructors have The push from above to for example train students on commercial software or have them be end users of software right that is Whereas I would have to I would like to teach coding as soon as possible and and even Python Why Python is is like reading a book sometimes right? We start with high-level languages Right and then we after they digest and understand high-level languages we go to low-level languages And then we can talk about high performance computing, but we're missing that preliminary step I believe to to train the students sometimes coding is considered the taboo Right they don't there should be I think every degree should have a Computer science class a hardcore computer science class as early as possible And I think some degrees are still lacking that so I would Argue that we need a step before that certainly think not every degree has to have a hardcore computer science Because I'm we don't we we can use these tools without having a computer science. I think you're programming to some degree An introductory programming course is interesting. I don't know that I don't think that everyone every engineer needs to be A programmer they certainly need to be a good expert user of the codes like finite elements here if you're a mechanical engineer You have to know how to program Folks in humanities that that there's programs that Combine programming and Algorithmic development with with humanitarian studies humanities. And so I think we can treat a coding Education as a more broad coding coding education before high-performance computing Computations there are courses of course offered, but it's not basically very directed curriculum and Including these advanced computing courses data analysis and all of those in sequence for a student get the training to get there How often all is needed for doing exit It doesn't mean necessarily that everyone Student to have good software engineering practices also right aggression testing Code repositories and nothing can replace actually hands-on writing code as you know mini projects to actually Do it in practice to get better? so as a consumer of our products you hire a lot of PhDs writing in your group over the years how do you see a difference in the Knowledge in the type of education that our students have over the over the years Do you see a an evolution of the type of background that then your staff members when they join? Sandhya have I look to students in that Have a lot of Interest in both the physical sciences as well as in computer science and computational science and algorithms So they kind of have a lot of them are hackers to begin with It happened to like physical science, too It's hard to find people like that with dual dual legs You know there are a number of universities that teach that kind of mixed curriculum And so for example at UT Austin has a very strong program in computational sciences Utah and there's a couple of places And where the students have exposure to computer science applied math and physical sciences and not just a stove pipe I Just this has been a fascinating discussion but I just have to come back down on the side of teaching the students computing and I would say really at the risk of being Controversial all engineering students should learn the program. It is interesting like 25 years ago There was a big swing in the pendulum where people were saying well Electrical engineers really shouldn't know how to program. It's not necessary because they'll be you know those be some sort of point-and-click Environment where they do this and it's gone exactly the opposite way more more companies and employers are just really want Students to be able to program because that's how you actually implement your ideas So and they want them to have good coding skills I Don't want we don't necessarily need all the students to do excellent scale competing. Oh, yeah, you know I took yours to be in line with my view Yeah, right. Yeah, you know, I totally agree. Yeah, but I don't think the coding education should be bypassed for user-friendly Exascale tool that maybe we can train them on should I think they should know what's under the hood