>>Oracle continues to enhance my sequel Heatwave at a very rapid pace. The company is now in its fourth major release since the original announcement in December 2020. 1 of the main criticisms of my sequel, Heatwave, is that it only runs on O. C I. Oracle Cloud Infrastructure and as a lock in to Oracle's Cloud. Oracle recently announced that heat wave is now going to be available in AWS Cloud and it announced its intent to bring my sequel Heatwave to Azure. So my secret heatwave on AWS is a significant TAM expansion move for Oracle because of the momentum AWS Cloud continues to show. And evidently the Heatwave Engineering team has taken the development effort from O. C I. And is bringing that to A W S with a number of enhancements that we're gonna dig into today is senior vice president. My sequel Heatwave at Oracle is back with me on a cube conversation to discuss the latest heatwave news, and we're eager to hear any benchmarks relative to a W S or any others. Nippon has been leading the Heatwave engineering team for over 10 years and there's over 100 and 85 patents and database technology. Welcome back to the show and good to see you. >>Thank you. Very happy to be back. >>Now for those who might not have kept up with the news, uh, to kick things off, give us an overview of my sequel, Heatwave and its evolution. So far, >>so my sequel, Heat Wave, is a fully managed my secret database service offering from Oracle. Traditionally, my secret has been designed and optimised for transaction processing. So customers of my sequel then they had to run analytics or when they had to run machine learning, they would extract the data out of my sequel into some other database for doing. Unlike processing or machine learning processing my sequel, Heat provides all these capabilities built in to a single database service, which is my sequel. He'd fake So customers of my sequel don't need to move the data out with the same database. They can run transaction processing and predicts mixed workloads, machine learning, all with a very, very good performance in very good price performance. Furthermore, one of the design points of heat wave is is a scale out architecture, so the system continues to scale and performed very well, even when customers have very large late assignments. >>So we've seen some interesting moves by Oracle lately. The collaboration with Azure we've we've covered that pretty extensively. What was the impetus here for bringing my sequel Heatwave onto the AWS cloud? What were the drivers that you considered? >>So one of the observations is that a very large percentage of users of my sequel Heatwave, our AWS users who are migrating of Aurora or so already we see that a good percentage of my secret history of customers are migrating from GWS. However, there are some AWS customers who are still not able to migrate the O. C. I to my secret heat wave. And the reason is because of, um, exorbitant cost, which was charges. So in order to migrate the workload from AWS to go see, I digress. Charges are very high fees which becomes prohibitive for the customer or the second example we have seen is that the latency of practising a database which is outside of AWS is very high. So there's a class of customers who would like to get the benefits of my secret heatwave but were unable to do so and with this support of my secret trip inside of AWS, these customers can now get all the grease of the benefits of my secret he trip without having to pay the high fees or without having to suffer with the poorly agency, which is because of the ws architecture. >>Okay, so you're basically meeting the customer's where they are. So was this a straightforward lifted shift from from Oracle Cloud Infrastructure to AWS? >>No, it is not because one of the design girls we have with my sequel, Heatwave is that we want to provide our customers with the best price performance regardless of the cloud. So when we decided to offer my sequel, he headed west. Um, we have optimised my sequel Heatwave on it as well. So one of the things to point out is that this is a service with the data plane control plane and the console are natively running on AWS. And the benefits of doing so is that now we can optimise my sequel Heatwave for the E. W s architecture. In addition to that, we have also announced a bunch of new capabilities as a part of the service which will also be available to the my secret history of customers and our CI, But we just announced them and we're offering them as a part of my secret history of offering on AWS. >>So I just want to make sure I understand that it's not like you just wrapped your stack in a container and stuck it into a W s to be hosted. You're saying you're actually taking advantage of the capabilities of the AWS cloud natively? And I think you've made some other enhancements as well that you're alluding to. Can you maybe, uh, elucidate on those? Sure. >>So for status, um, we have taken the mind sequel Heatwave code and we have optimised for the It was infrastructure with its computer network. And as a result, customers get very good performance and price performance. Uh, with my secret he trade in AWS. That's one performance. Second thing is, we have designed new interactive counsel for the service, which means that customers can now provision there instances with the council. But in addition, they can also manage their schemas. They can. Then court is directly from the council. Autopilot is integrated. The council we have introduced performance monitoring, so a lot of capabilities which we have introduced as a part of the new counsel. The third thing is that we have added a bunch of new security features, uh, expose some of the security features which were part of the My Secret Enterprise edition as a part of the service, which gives customers now a choice of using these features to build more secure applications. And finally, we have extended my secret autopilot for a number of old gpus cases. In the past, my secret autopilot had a lot of capabilities for Benedict, and now we have augmented my secret autopilot to offer capabilities for elderly people. Includes as well. >>But there was something in your press release called Auto thread. Pooling says it provides higher and sustained throughput. High concerns concerns concurrency by determining Apple number of transactions, which should be executed. Uh, what is that all about? The auto thread pool? It seems pretty interesting. How does it affect performance? Can you help us understand that? >>Yes, and this is one of the capabilities of alluding to which we have added in my secret autopilot for transaction processing. So here is the basic idea. If you have a system where there's a large number of old EP transactions coming into it at a high degrees of concurrency in many of the existing systems of my sequel based systems, it can lead to a state where there are few transactions executing, but a bunch of them can get blocked with or a pilot tried pulling. What we basically do is we do workload aware admission control and what this does is it figures out, what's the right scheduling or all of these algorithms, so that either the transactions are executing or as soon as something frees up, they can start executing, so there's no transaction which is blocked. The advantage to the customer of this capability is twofold. A get significantly better throughput compared to service like Aurora at high levels of concurrency. So at high concurrency, for instance, uh, my secret because of this capability Uh oh, thread pulling offers up to 10 times higher compared to Aurora, that's one first benefit better throughput. The second advantage is that the true part of the system never drops, even at high levels of concurrency, whereas in the case of Aurora, the trooper goes up, but then, at high concurrency is, let's say, starting, uh, level of 500 or something. It depends upon the underlying shit they're using the troopers just dropping where it's with my secret heatwave. The truth will never drops. Now, the ramification for the customer is that if the truth is not gonna drop, the user can start off with a small shape, get the performance and be a show that even the workload increases. They will never get a performance, which is worse than what they're getting with lower levels of concurrency. So this let's leads to customers provisioning a shape which is just right for them. And if they need, they can, uh, go with the largest shape. But they don't like, you know, over pay. So those are the two benefits. Better performance and sustain, uh, regardless of the level of concurrency. >>So how do we quantify that? I know you've got some benchmarks. How can you share comparisons with other cloud databases especially interested in in Amazon's own databases are obviously very popular, and and are you publishing those again and get hub, as you have done in the past? Take us through the benchmarks. >>Sure, So benchmarks are important because that gives customers a sense of what performance to expect and what price performance to expect. So we have run a number of benchmarks. And yes, all these benchmarks are available on guitar for customers to take a look at. So we have performance results on all the three castle workloads, ol DB Analytics and Machine Learning. So let's start with the Rdp for Rdp and primarily because of the auto thread pulling feature. We show that for the IPCC for attended dataset at high levels of concurrency, heatwave offers up to 10 times better throughput and this performance is sustained, whereas in the case of Aurora, the performance really drops. So that's the first thing that, uh, tend to alibi. Sorry, 10 gigabytes. B B C c. I can come and see the performance are the throughput is 10 times better than Aurora for analytics. We have done a comparison of my secret heatwave in AWS and compared with Red Ship Snowflake Googled inquiry, we find that the price performance of my secret heatwave compared to read ship is seven times better. So my sequel, Heat Wave in AWS, provides seven times better price performance than red ship. That's a very, uh, interesting results to us. Which means that customers of Red Shift are really going to take the service seriously because they're gonna get seven times better price performance. And this is all running in a W s so compared. >>Okay, carry on. >>And then I was gonna say, compared to like, Snowflake, uh, in AWS offers 10 times better price performance. And compared to Google, ubiquity offers 12 times better price performance. And this is based on a four terabyte p PCH workload. Results are available on guitar, and then the third category is machine learning and for machine learning, uh, for training, the performance of my secret heatwave is 25 times faster compared to that shit. So all the three workloads we have benchmark's results, and all of these scripts are available on YouTube. >>Okay, so you're comparing, uh, my sequel Heatwave on AWS to Red Shift and snowflake on AWS. And you're comparing my sequel Heatwave on a W s too big query. Obviously running on on Google. Um, you know, one of the things Oracle is done in the past when you get the price performance and I've always tried to call fouls you're, like, double your price for running the oracle database. Uh, not Heatwave, but Oracle Database on a W s. And then you'll show how it's it's so much cheaper on on Oracle will be like Okay, come on. But they're not doing that here. You're basically taking my sequel Heatwave on a W s. I presume you're using the same pricing for whatever you see to whatever else you're using. Storage, um, reserved instances. That's apples to apples on A W s. And you have to obviously do some kind of mapping for for Google, for big query. Can you just verify that for me, >>we are being more than fair on two dimensions. The first thing is, when I'm talking about the price performance for analytics, right for, uh, with my secret heat rape, the cost I'm talking about from my secret heat rape is the cost of running transaction processing, analytics and machine learning. So it's a fully loaded cost for the case of my secret heatwave. There has been I'm talking about red ship when I'm talking about Snowflake. I'm just talking about the cost of these databases for running, and it's only it's not, including the source database, which may be more or some other database, right? So that's the first aspect that far, uh, trip. It's the cost for running all three kinds of workloads, whereas for the competition, it's only for running analytics. The second thing is that for these are those services whether it's like shit or snowflakes, That's right. We're talking about one year, fully paid up front cost, right? So that's what most of the customers would pay for. Many of the customers would pay that they will sign a one year contract and pay all the costs ahead of time because they get a discount. So we're using that price and the case of Snowflake. The costs were using is their standard edition of price, not the Enterprise edition price. So yes, uh, more than in this competitive. >>Yeah, I think that's an important point. I saw an analysis by Marx Tamer on Wiki Bond, where he was doing the TCO comparisons. And I mean, if you have to use two separate databases in two separate licences and you have to do et yelling and all the labour associated with that, that that's that's a big deal and you're not even including that aspect in in your comparison. So that's pretty impressive. To what do you attribute that? You know, given that unlike, oh, ci within the AWS cloud, you don't have as much control over the underlying hardware. >>So look hard, but is one aspect. Okay, so there are three things which give us this advantage. The first thing is, uh, we have designed hateful foreign scale out architecture. So we came up with new algorithms we have come up with, like, uh, one of the design points for heat wave is a massively partitioned architecture, which leads to a very high degree of parallelism. So that's a lot of hype. Each were built, So that's the first part. The second thing is that although we don't have control over the hardware, but the second design point for heat wave is that it is optimised for commodity cloud and the commodity infrastructure so we can have another guys, what to say? The computer we get, how much network bandwidth do we get? How much of, like objects to a brand that we get in here? W s. And we have tuned heat for that. That's the second point And the third thing is my secret autopilot, which provides machine learning based automation. So what it does is that has the users workload is running. It learns from it, it improves, uh, various premieres in the system. So the system keeps getting better as you learn more and more questions. And this is the third thing, uh, as a result of which we get a significant edge over the competition. >>Interesting. I mean, look, any I SV can go on any cloud and take advantage of it. And that's, uh I love it. We live in a new world. How about machine learning workloads? What? What did you see there in terms of performance and benchmarks? >>Right. So machine learning. We offer three capabilities training, which is fully automated, running in France and explanations. So one of the things which many of our customers told us coming from the enterprise is that explanations are very important to them because, uh, customers want to know that. Why did the the system, uh, choose a certain prediction? So we offer explanations for all models which have been derailed by. That's the first thing. Now, one of the interesting things about training is that training is usually the most expensive phase of machine learning. So we have spent a lot of time improving the performance of training. So we have a bunch of techniques which we have developed inside of Oracle to improve the training process. For instance, we have, uh, metal and proxy models, which really give us an advantage. We use adaptive sampling. We have, uh, invented in techniques for paralysing the hyper parameter search. So as a result of a lot of this work, our training is about 25 times faster than that ship them health and all the data is, uh, inside the database. All this processing is being done inside the database, so it's much faster. It is inside the database. And I want to point out that there is no additional charge for the history of customers because we're using the same cluster. You're not working in your service. So all of these machine learning capabilities are being offered at no additional charge inside the database and as a performance, which is significantly faster than that, >>are you taking advantage of or is there any, uh, need not need, but any advantage that you can get if two by exploiting things like gravity. John, we've talked about that a little bit in the past. Or trainee. Um, you just mentioned training so custom silicon that AWS is doing, you're taking advantage of that. Do you need to? Can you give us some insight >>there? So there are two things, right? We're always evaluating What are the choices we have from hybrid perspective? Obviously, for us to leverage is right and like all the things you mention about like we have considered them. But there are two things to consider. One is he is a memory system. So he favours a big is the dominant cost. The processor is a person of the cost, but memory is the dominant cost. So what we have evaluated and found is that the current shape which we are using is going to provide our customers with the best price performance. That's the first thing. The second thing is that there are opportunities at times when we can use a specialised processor for vaccinating the world for a bit. But then it becomes a matter of the cost of the customer. Advantage of our current architecture is on the same hardware. Customers are getting very good performance. Very good, energetic performance in a very good machine learning performance. If you will go with the specialised processor, it may. Actually, it's a machine learning, but then it's an additional cost with the customers we need to pay. So we are very sensitive to the customer's request, which is usually to provide very good performance at a very low cost. And we feel is that the current design we have as providing customers very good performance and very good price performance. >>So part of that is architectural. The memory intensive nature of of heat wave. The other is A W s pricing. If AWS pricing were to flip, it might make more sense for you to take advantage of something like like cranium. Okay, great. Thank you. And welcome back to the benchmarks benchmarks. Sometimes they're artificial right there. A car can go from 0 to 60 in two seconds. But I might not be able to experience that level of performance. Do you? Do you have any real world numbers from customers that have used my sequel Heatwave on A W s. And how they look at performance? >>Yes, absolutely so the my Secret service on the AWS. This has been in Vera for, like, since November, right? So we have a lot of customers who have tried the service. And what actually we have found is that many of these customers, um, planning to migrate from Aurora to my secret heat rape. And what they find is that the performance difference is actually much more pronounced than what I was talking about. Because with Aurora, the performance is actually much poorer compared to uh, like what I've talked about. So in some of these cases, the customers found improvement from 60 times, 240 times, right? So he travels 100 for 240 times faster. It was much less expensive. And the third thing, which is you know, a noteworthy is that customers don't need to change their applications. So if you ask the top three reasons why customers are migrating, it's because of this. No change to the application much faster, and it is cheaper. So in some cases, like Johnny Bites, what they found is that the performance of their applications for the complex storeys was about 60 to 90 times faster. Then we had 60 technologies. What they found is that the performance of heat we have compared to Aurora was 100 and 39 times faster. So, yes, we do have many such examples from real workloads from customers who have tried it. And all across what we find is if it offers better performance, lower cost and a single database such that it is compatible with all existing by sequel based applications and workloads. >>Really impressive. The analysts I talked to, they're all gaga over heatwave, and I can see why. Okay, last question. Maybe maybe two and one. Uh, what's next? In terms of new capabilities that customers are going to be able to leverage and any other clouds that you're thinking about? We talked about that upfront, but >>so in terms of the capabilities you have seen, like they have been, you know, non stop attending to the feedback from the customers in reacting to it. And also, we have been in a wedding like organically. So that's something which is gonna continue. So, yes, you can fully expect that people not dressed and continue to in a way and with respect to the other clouds. Yes, we are planning to support my sequel. He tripped on a show, and this is something that will be announced in the near future. Great. >>All right, Thank you. Really appreciate the the overview. Congratulations on the work. Really exciting news that you're moving my sequel Heatwave into other clouds. It's something that we've been expecting for some time. So it's great to see you guys, uh, making that move, and as always, great to have you on the Cube. >>Thank you for the opportunity. >>All right. And thank you for watching this special cube conversation. I'm Dave Volonte, and we'll see you next time.

>>welcome. My name is Ben Bennett. I am the director of HPC Strategic programs here at Hewlett Packard Enterprise. It is my great pleasure and honor to be talking to Professor Mark Parsons from the Edinburgh Parallel Computing Center. And we're gonna talk a little about exa scale. What? It means we're gonna talk less about the technology on Maura about the science, the requirements on the need for exa scale. Uh, rather than a deep dive into the enabling technologies. Mark. Welcome. >>I then thanks very much for inviting me to tell me >>complete pleasure. Um, so I'd like to kick off with, I suppose. Quite an interesting look back. You and I are both of a certain age 25 plus, Onda. We've seen these milestones. Uh, I suppose that the S I milestones of high performance computing's come and go, you know, from a gig a flop back in 1987 teraflop in 97 a petaflop in 2000 and eight. But we seem to be taking longer in getting to an ex a flop. Um, so I'd like your thoughts. Why is why is an extra flop taking so long? >>So I think that's a very interesting question because I started my career in parallel computing in 1989. I'm gonna join in. IPCC was set up then. You know, we're 30 years old this year in 1990 on Do you know the fastest computer we have them is 800 mega flops just under a getting flogged. So in my career, we've gone already. When we reached the better scale, we'd already gone pretty much a million times faster on, you know, the step from a tariff block to a block scale system really didn't feel particularly difficult. Um, on yet the step from A from a petaflop PETA scale system. To an extent, block is a really, really big challenge. And I think it's really actually related to what's happened with computer processes over the last decade, where, individually, you know, approached the core, Like on your laptop. Whoever hasn't got much faster, we've just got more often So the perception of more speed, but actually just being delivered by more course. And as you go down that approach, you know what happens in the supercomputing world as well. We've gone, uh, in 2010 I think we had systems that were, you know, a few 1000 cores. Our main national service in the UK for the last eight years has had 118,000 cores. But looking at the X scale we're looking at, you know, four or five million cores on taming that level of parallelism is the real challenge. And that's why it's taking an enormous and time to, uh, deliver these systems. That is not just on the hardware front. You know, vendors like HP have to deliver world beating technology and it's hard, hard. But then there's also the challenge to the users. How do they get the codes to work in the face of that much parallelism? >>If you look at what the the complexity is delivering an annex a flop. Andi, you could have bought an extra flop three or four years ago. You couldn't have housed it. You couldn't have powered it. You couldn't have afforded it on, do you? Couldn't program it. But you still you could have You could have bought one. We should have been so lucky to be unable to supply it. Um, the software, um I think from our standpoint, is is looking like where we're doing mawr enabling with our customers. You sell them a machine on, then the the need then to do collaboration specifically seems mawr and Maura around the software. Um, so it's It's gonna be relatively easy to get one x a flop using limb pack, but but that's not extra scale. So what do you think? On exa scale machine versus an X? A flop machine means to the people like yourself to your users, the scientists and industry. What is an ex? A flop versus >>an exa scale? So I think, you know, supercomputing moves forward by setting itself challenges. And when you when you look at all of the excess scale programs worldwide that are trying to deliver systems that can do an X a lot form or it's actually very arbitrary challenge. You know, we set ourselves a PETA scale challenge delivering a petaflop somebody manage that, Andi. But you know, the world moves forward by setting itself challenges e think you know, we use quite arbitrary definition of what we mean is well by an exit block. So, you know, in your in my world, um, we either way, first of all, see ah flop is a computation, so multiply or it's an ad or whatever on we tend. Thio, look at that is using very high precision numbers or 64 bit numbers on Do you know, we then say, Well, you've got to do the next block. You've got to do a billion billion of those calculations every second. No, a some of the last arbitrary target Now you know today from HPD Aiken by my assistant and will do a billion billion calculations per second. And they will either do that as a theoretical peak, which would be almost unattainable, or using benchmarks that stressed the system on demonstrate a relaxing law. But again, those benchmarks themselves attuned Thio. Just do those calculations and deliver and explore been a steady I'll way if you like. So, you know, way kind of set ourselves this this this big challenge You know, the big fence on the race course, which were clambering over. But the challenge in itself actually should be. I'm much more interesting. The water we're going to use these devices for having built um, eso. Getting into the extra scale era is not so much about doing an extra block. It's a new generation off capability that allows us to do better scientific and industrial research. And that's the interesting bit in this whole story. >>I would tend to agree with you. I think the the focus around exa scale is to look at, you know, new technologies, new ways of doing things, new ways of looking at data and to get new results. So eventually you will get yourself a nexus scale machine. Um, one hopes, sooner rather >>than later. Well, I'm sure you don't tell me one, Ben. >>It's got nothing to do with may. I can't sell you anything, Mark. But there are people outside the door over there who would love to sell you one. Yes. However, if we if you look at your you know your your exa scale machine, Um, how do you believe the workloads are going to be different on an extra scale machine versus your current PETA scale machine? >>So I think there's always a slight conceit when you buy a new national supercomputer. On that conceit is that you're buying a capability that you know on. But many people will run on the whole system. Known truth. We do have people that run on the whole of our archer system. Today's A 118,000 cores, but I would say, and I'm looking at the system. People that run over say, half of that can be counted on Europe on a single hand in a year, and they're doing very specific things. It's very costly simulation they're running on. So, you know, if you look at these systems today, two things show no one is. It's very difficult to get time on them. The Baroque application procedures All of the requirements have to be assessed by your peers and your given quite limited amount of time that you have to eke out to do science. Andi people tend to run their applications in the sweet spot where their application delivers the best performance on You know, we try to push our users over time. Thio use reasonably sized jobs. I think our average job says about 20,000 course, she's not bad, but that does mean that as we move to the exits, kill two things have to happen. One is actually I think we've got to be more relaxed about giving people access to the system, So let's give more people access, let people play, let people try out ideas they've never tried out before. And I think that will lead to a lot more innovation and computational science. But at the same time, I think we also need to be less precious. You know, we to accept these systems will have a variety of sizes of job on them. You know, we're still gonna have people that want to run four million cores or two million cores. That's absolutely fine. Absolutely. Salute those people for trying really, really difficult. But then we're gonna have a huge spectrum of views all the way down to people that want to run on 500 cores or whatever. So I think we need Thio broaden the user base in Alexa Skill system. And I know this is what's happening, for example, in Japan with the new Japanese system. >>So, Mark, if you cast your mind back to almost exactly a year ago after the HPC user forum, you were interviewed for Premier Magazine on Do you alluded in that article to the needs off scientific industrial users requiring, you know, uh on X a flop or an exa scale machine it's clear in your in your previous answer regarding, you know, the workloads. Some would say that the majority of people would be happier with, say, 10 100 petaflop machines. You know, democratization. More people access. But can you provide us examples at the type of science? The needs of industrial users that actually do require those resources to be put >>together as an exa scale machine? So I think you know, it's a very interesting area. At the end of the day, these systems air bought because they are capability systems on. I absolutely take the argument. Why shouldn't we buy 10 100 pattern block systems? But there are a number of scientific areas even today that would benefit from a nexus school system and on these the sort of scientific areas that will use as much access onto a system as much time and as much scale of the system as they can, as you can give them eso on immediate example. People doing chroma dynamics calculations in particle physics, theoretical calculations, they would just use whatever you give them. But you know, I think one of the areas that is very interesting is actually the engineering space where, you know, many people worry the engineering applications over the last decade haven't really kept up with this sort of supercomputers that we have. I'm leading a project called Asimov, funded by M. P S O. C in the UK, which is jointly with Rolls Royce, jointly funded by Rolls Royce and also working with the University of Cambridge, Oxford, Bristol, Warrick. We're trying to do the whole engine gas turbine simulation for the first time. So that's looking at the structure of the gas turbine, the airplane engine, the structure of it, how it's all built it together, looking at the fluid dynamics off the air and the hot gasses, the flu threat, looking at the combustion of the engine looking how fuel is spread into the combustion chamber. Looking at the electrics around, looking at the way the engine two forms is, it heats up and cools down all of that. Now Rolls Royce wants to do that for 20 years. Andi, Uh, whenever they certify, a new engine has to go through a number of physical tests, and every time they do on those tests, it could cost them as much as 25 to $30 million. These are very expensive tests, particularly when they do what's called a blade off test, which would be, you know, blade failure. They could prove that the engine contains the fragments of the blade. Sort of think, continue face really important test and all engines and pass it. What we want to do is do is use an exa scale computer to properly model a blade off test for the first time, so that in future, some simulations can become virtual rather than having thio expend all of the money that Rolls Royce would normally spend on. You know, it's a fascinating project is a really hard project to do. One of the things that I do is I am deaf to share this year. Gordon Bell Price on bond I've really enjoyed to do. That's one of the major prizes in our area, you know, gets announced supercomputing every year. So I have the pleasure of reading all the submissions each year. I what's been really interesting thing? This is my third year doing being on the committee on what's really interesting is the way that big systems like Summit, for example, in the US have pushed the user communities to try and do simulations Nowhere. Nobody's done before, you know. And we've seen this as well, with papers coming after the first use of the for Goku system in Japan, for example, people you know, these are very, very broad. So, you know, earthquake simulation, a large Eddie simulations of boats. You know, a number of things around Genome Wide Association studies, for example. So the use of these computers spans of last area off computational science. I think the really really important thing about these systems is their challenging people that do calculations they've never done before. That's what's important. >>Okay, Thank you. You talked about challenges when I nearly said when you and I had lots of hair, but that's probably much more true of May. Um, we used to talk about grand challenges we talked about, especially around the teraflop era, the ski red program driving, you know, the grand challenges of science, possibly to hide the fact that it was a bomb designing computer eso they talked about the grand challenges. Um, we don't seem to talk about that much. We talk about excess girl. We talk about data. Um Where are the grand challenges that you see that an exa scale computer can you know it can help us. Okay, >>so I think grand challenges didn't go away. Just the phrase went out of fashion. Um, that's like my hair. I think it's interesting. The I do feel the science moves forward by setting itself grand challenges and always had has done, you know, my original backgrounds in particle physics. I was very lucky to spend four years at CERN working in the early stage of the left accelerator when it first came online on. Do you know the scientists there? I think they worked on left 15 years before I came in and did my little ph d on it. Andi, I think that way of organizing science hasn't changed. We just talked less about grand challenges. I think you know what I've seen over the last few years is a renaissance in computational science, looking at things that have previously, you know, people have said have been impossible. So a couple of years ago, for example, one of the key Gordon Bell price papers was on Genome Wide Association studies on some of it. If I may be one of the winner of its, if I remember right on. But that was really, really interesting because first of all, you know, the sort of the Genome Wide Association Studies had gone out of favor in the bioinformatics by a scientist community because people thought they weren't possible to compute. But that particular paper should Yes, you could do these really, really big Continental little problems in a reasonable amount of time if you had a big enough computer. And one thing I felt all the way through my career actually is we've probably discarded Mawr simulations because they were impossible at the time that we've actually decided to do. And I sometimes think we to challenge ourselves by looking at the things we've discovered in the past and say, Oh, look, you know, we could actually do that now, Andi, I think part of the the challenge of bringing an extra service toe life is to get people to think about what they would use it for. That's a key thing. Otherwise, I always say, a computer that is unused to just be turned off. There's no point in having underutilized supercomputer. Everybody loses from that. >>So Let's let's bring ourselves slightly more up to date. We're in the middle of a global pandemic. Uh, on board one of the things in our industry has bean that I've been particularly proud about is I've seen the vendors, all the vendors, you know, offering up machine's onboard, uh, making resources available for people to fight things current disease. Um, how do you see supercomputers now and in the future? Speeding up things like vaccine discovery on help when helping doctors generally. >>So I think you're quite right that, you know, the supercomputer community around the world actually did a really good job of responding to over 19. Inasmuch as you know, speaking for the UK, we put in place a rapid access program. So anybody wanted to do covert research on the various national services we have done to the to two services Could get really quick access. Um, on that, that has worked really well in the UK You know, we didn't have an archer is an old system, Aziz. You know, we didn't have the world's largest supercomputer, but it is happily bean running lots off covert 19 simulations largely for the biomedical community. Looking at Druk modeling and molecular modeling. Largely that's just been going the US They've been doing really large uh, combinatorial parameter search problems on on Summit, for example, looking to see whether or not old drugs could be reused to solve a new problem on DSO, I think, I think actually, in some respects Kobe, 19 is being the sounds wrong. But it's actually been good for supercomputing. Inasmuch is pointed out to governments that supercomputers are important parts off any scientific, the active countries research infrastructure. >>So, um, I'll finish up and tap into your inner geek. Um, there's a lot of technologies that are being banded around to currently enable, you know, the first exa scale machine, wherever that's going to be from whomever, what are the current technologies or emerging technologies that you are interested in excited about looking forward to getting your hands on. >>So in the business case I've written for the U. K's exa scale computer, I actually characterized this is a choice between the American model in the Japanese model. Okay, both of frozen, both of condoms. Eso in America, they're very much gone down the chorus plus GPU or GPU fruit. Um, so you might have, you know, an Intel Xeon or an M D process er center or unarmed process or, for that matter on you might have, you know, 24 g. P. U s. I think the most interesting thing that I've seen is definitely this move to a single address space. So the data that you have will be accessible, but the G p u on the CPU, I think you know, that's really bean. One of the key things that stopped the uptake of GPS today and that that that one single change is going Thio, I think, uh, make things very, very interesting. But I'm not entirely convinced that the CPU GPU model because I think that it's very difficult to get all the all the performance set of the GPU. You know, it will do well in H p l, for example, high performance impact benchmark we're discussing at the beginning of this interview. But in riel scientific workloads, you know, you still find it difficult to find all the performance that has promised. So, you know, the Japanese approach, which is the core, is only approach. E think it's very attractive, inasmuch as you know They're using very high bandwidth memory, very interesting process of which they are going to have to, you know, which they could develop together over 10 year period. And this is one thing that people don't realize the Japanese program and the American Mexico program has been working for 10 years on these systems. I think the Japanese process really interesting because, um, it when you look at the performance, it really does work for their scientific work clothes, and that's that does interest me a lot. This this combination of a A process are designed to do good science, high bandwidth memory and a real understanding of how data flows around the supercomputer. I think those are the things are exciting me at the moment. Obviously, you know, there's new networking technologies, I think, in the fullness of time, not necessarily for the first systems. You know, over the next decade we're going to see much, much more activity on silicon photonics. I think that's really, really fascinating all of these things. I think in some respects the last decade has just bean quite incremental improvements. But I think we're supercomputing is going in the moment. We're a very very disruptive moment again. That goes back to start this discussion. Why is extra skill been difficult to get? Thio? Actually, because the disruptive moment in technology. >>Professor Parsons, thank you very much for your time and your insights. Thank you. Pleasure and folks. Thank you for watching. I hope you've learned something, or at least enjoyed it. With that, I would ask you to stay safe and goodbye.