The calculation of the calculation that is developed in a way of exponential in the last period is the ability to solve problems, many times already known at the theoretical level, but not resolvable in time, in the way of using the calculation. We use the calculation of the energy sector for the fluid dynamics, for the research of materials, for the analysis molecular, for all the studies that are also of evolution, for example, the catture, the stockage of CO2, the model of the field of the sea, which is offshore or offshore, and as well as the geosciences, which are the ones that have always been characterized. In these areas, having the capacity of calculating, having the competence inside and being able to develop software that is useful at the maximum capacity of calculating parallel is fundamental and crucial to solve problems that otherwise would remain solved. Welcome to Emiliano Audizio from Wired Italia and this is a new episode of The Big Interview Live from Bocconi. This episode is made in collaboration with Eni and in this episode we will talk about super computers, super calculus, digital infrastructure, energy and we will do it with Dario Pagani Head of Digital Information Technology of Eni. Welcome Dario. Thank you Emiliano. We have said to give you the tool so I am happy with you. Very nice. Parting from here. You have inaugurated in province of Pavia HPC6, which is a computer that is a level industrial, the most powerful in the world. I want to ask you what are the characteristics of this computer and also what is a computer? We have been in the years that we invest in supercalculation, so that HPC6 is the sixth generation of super calculator. In super calculator, in hardware, like the HPC6, but also in the development of software, software parallel, which can use the characteristics of a supercomputer, to solve problems that we have in the sector energetically. HPC6 has been an important step in terms of calculating the calculator, because we have in the current installation that we have, HPC4 and HPC5, have circa una capacità di calcolo di 70 petaflop HPC6 a 606 petaflop di picco. Ma cosa sono questi petaflop? Sono nient'altro che milioni di miliardi di operazioni al secondo complesse. Milioni di miliardi di operazioni al secondo vuol dire fare grandi calcoli, grande velocità, grande capacità di risolvere, poi equazioni complesse, modelli complessi, equazioni matematiche, problemi fisici, which gave us the ability, through software and supercalculated, to be recognized as leaders, especially in the geosciences, in the study of natural energy. HPC6 is a calculator installed in our data center in the province of Pavia. And, while we're talking about this, the need for super-calculation, so that we have a great laborator, a great calculation, if we think about the research and the energy, and you have done so much research, today, if we think about the last two Nobel Prize for physics and chemistry, without a great calculation, that would be great? Absolutely. The calculation of the calculation that is developed in a way of exponential in the last period is the ability to solve problems. Many times already known at the theoretical level, but not solve them in time, in the way of using the calculation. We use the calculation of the calculation, as I said, in the energy sector for the fluid dynamics, for the research materials, for analysis molecular, for all the studies that are also of evolution, for example, on the catture, of the stockage of CO2, the model of the field of the sea, which is offshore or offshore, and on the geosciences, which are the ones that have always been characterized. In these areas, having the capacity of calculating, having the competence inside and being able to develop software that is useful at the maximum capacity of calculating parallel is fundamental and crucial to solve problems that otherwise remain solved. You mentioned the research issue, we, especially lately, are accelerating a lot of research, a lot of research with super calculus and also with a way of using the artificial intelligence not so traditional as we read in a natural way, in a way, if we look at the artificial intelligence, this, which we use in some types of models for a long time, allows us to accelerate the research phase in order to be able to see immediately if what you are looking for can escalate, can be profitable, can be implemented, without expecting the typical sequential approach, I research, I do, I do the first pilot and then I can see. This gives us a great competitive advantage and especially if we want to risk many and our internal process. This element is also very important from a point of view of development, in the sense that you are a manager, and you have many years that you take care of these topics. You said earlier that we have to encourage the time so if you want to remain competitive it becomes necessary to be able to use tools that to accorciare i tempi quindi se si vuole rimanere competitivi diventa necessario riuscire a dotarsi di strumenti che rispondono a questa esigenza Assolutamente, nel senso che io dico sempre che ad oggi abbiamo una grande opportunità che termina dal fatto della disponibilità di tecnologia, però chiaramente la tecnologia non è nulla se poi tu non riesci ovviamente a utilizzarla e qui mettiamo chiaramente la centralità of our people, which are the real value in the company, the people, the competencies, the experience to put in common, to use the technology that is a tool, a tool that can accelerate, that can help you, that can differenziate your ability to stay in business, to stay in the market, to stay as an organization to be competitive. But it is an important instrument to be able to use, but the centricity of the skills that you mature in business with our people, with the story of our people and continually alimenting these skills is the key. From this point of view, you said it well, I think it's very important to me that you want to touch the topic of the instrument, you have talked about the supercalculation, and an element, because another key word of this discussion is infrastructure digital. It's also a couple of words that we often hear, but sometimes it's difficult to understand the origin, HPC6 is inserted into an digital infrastructure called Green Data Center. Tell us about digital infrastructure, how it is, and how it is evolved from a point of view energetically. Green Data Center is one of our vision. The data center was already completed 10 years ago, but it was thought about 12-13 years ago. It seems to be a bit of an acronistic idea, since we have access to infrastructure in a way more easily. For a simple reason, to put a clock in 10 years, just to put it in that time, there were no investors in the digital infrastructure in Italy that could be useful for us. and to build a infrastructure, I call it a house of computers, we don't have the HPC, but we have the other traditional systems, that serve the business, but you can't have a development strategy if you don't have a house that can be hospitable, efficient and sustainable, And at the time, the attribute green says, because we thought it was super efficient. A data center is an energy source, because the calculation is not enough to transform energy in a calculation and dissipate the heat that emet is what also do our computer. At that time we said that we would do one that is super efficient. Ten in conto che a quell'epoca la media del rapporto di efficienza in data center era intorno ai 3, vuol dire che per ogni kilowatt di calcolo ne servivano 3 di poter alimentare tra alimentare, dissipazione e tutto quello che serve come consumi. We have a green one that had a rapport of 1,12%. That means that the 12% of energy that didn't go to the calculation. So it's super efficient to increase the energy, but above all the emissions of CO2. Because if you have a concept that we have, if we want, a little bit from the story, a little bit from the ancient Greece, which are called the free cooling, which is called the heat of a force, through the camines, through large volumes, which automatically makes it possible to use to the computer, to the computer, this type of heat. To make a computer efficient, capable of hospitable these machines in a contemporaneity, we said we have HPC6, but we have HPC4, HPC5. Without continuity solution, our business continues to work while we, with about 3 or 4 years, we can't do super-calculate. It was essential. If you don't have an infrastructure of this type, you can't do it. Our vision was correct, our strategy was correct and it was giving us great satisfaction, even now, for this concept of infrastructure capable of hospitality, digital infrastructure, such as the facility of a data centre, but also the super-calculated infrastructure, and, as a metaphor, not usual, is a topic of making a tuff. We have to have a trampoline that is just elastic, that can be sensible, that can be capable of expressing the tuffer. Then we do the tuff we do. In this sense, can you tell us a concrete case where this infrastructure, this calculation, which is thought and also designed and designed in order to evolve in time, according to performance performance, but also to efficiency, which helped the development of some products, something that we use in life and in every day. Well, I can tell you that now, for example, I'll just add a little bit. The HPC6, in evolution technological, is no longer a computer that is possible to heat it up to a water, but it is necessary to heat it up to a liquid. We have adapted our data center to heat it up to the 96% of the heat through the water. This is an important tension to be if we want to be sustainable to optimize everything So the new system for density is about 10 KW per square meter which is a huge power we are almost at the maximum density of computer science and we have a liquid Can I give you an example of this supercalculation that goes over the business? We have used our supercalculation in the Covid-19, for doing research on the molecules of the medicine that could have great success to fight the coronavirus. We understand that our competencies on molecular issues, but not applied to the business of energy, but in this case a great database of the Don Pei's properties, in this case, on molecules, has allowed to find a medical medicine that had excellent capacity to fight the coronavirus. This is a great pride, because it means that having computer skills, having capacity and competencies, and to make contaminations with other sectors can make great success. And here I would like to add, since these instruments and infrastructure allow steps that were not possible, but the collaboration and the aperture that were not even imaginable. I would like to ask that, since we talk about artificial intelligence, of the digital world, there is always a question related to people, as well as the people who work, how they are also attached to their their own professional method. All right, I ask you, how do you manage this transformation from the point of view of the competencies and the people who are involved in your production process? We have sposated and we believe that the intelligence artificial, in general and complex, but even now, the potential of intelligence artificial, has a very clear model, having the human center. So, to be a instrument, we have a service of human, that can be a service of human, which can be the skills and skills, but that allows the decision-making phase, the decision-making phase, in the moment of inferencing, where we are going to azzerate the bias, which can be a very important topic, as in the moment of decision-making. So it's a supervision with the human in the loop, the human in the loop, the human in the loop, These new instruments require also an evolution of the mindset of our people in understanding that this is a strong instrument that can be strengthened and not a strong instrument that I delego completely to some type of operations. Here the risk that we can see is that of the loss of our cognitive capacity, of our ability to make the difference as human beings. This is why it is always more potent, in addition to the technical competence, also the competence of vision, the competence of idea, the competence of an added value. So the internet artificial is accompanied, process by process, and is made by the responsible discussion. Today we have also a IAC to the European and Italian, which gives certain types of indications. This, if we turn on, not in an adempment, but in a model with which you can grow and leave the company and leave the expertise and the people that are the real value of the company in time, this is the model that is vincente. The big interview is also the key to try to stimulate the idea of the critical thinking, especially today, I refer to the artificial intelligence. I can't, this is my opinion, to prescindere from a tentative of study of the mechanism of function, of the logic of the function, because this logic, which is a word that comes from philosophy, So can we understand how to interact with machines and how our vision can benefit from the power of computing machines? Absolutely, as I said, we have been discussing a debate in some areas, there is also a report of MIT that says that today the topic is how to manage cognitive capabilities. It's clear that today we need to be more in the course of the teaching and the preparation of people, the development and continuous learning of people, to be able to not only do the technical skills, which are important, but all those that you said, all those that we have as human beings, so that one of the debates is that the artificial intelligence is all intelligent. Also on the topic, there is a beautiful story that we will tell you now, even how the term artificial intelligence is, I invite you to research that is not even a case of history, not a lot of it. Many researchers are contrary to this definition. Absolutely, yes, I agree with them. It is a great opportunity, it is a great opportunity, it is a technology that should always hold the human being, Of course, human beings are capable of being the distinct factor in all the activities. There is still more need to study, to develop, to develop. I always say, as usual, I don't want to lose cognitive skills. This happens also through generations, it can be banal. Today, many new generations have lost the cognitive ability of orienting. because we are delegating to someone else, or memorizing of numbers or simple things I think there are also more complex things When I delegating something I aware that I delegating something that this is probably a one way process a process of coming out, but that is difficult to recover if I need it. And then it's clear that I'm not totally dependent, but I have an autonomy, an autonomy, a power of thinking, which then power the technology. But not only. My mom said that in the world there are no more disceses, because after all the disceses are the salutes. Well, a good advice. I would like to talk about two topics. The first thing you mentioned before, when you started to project this supercalculation model, this infrastructure of supercalculation, there was an offer on the market that was different from the current. So, we have to say that certain capabilities and certain capabilities were not immediately available. Today there are great world players that offer these capabilities. I wonder how you decided to continue to develop a proprietary model. For a simple reason, when this model allows you to be recognized as an excellent model, so it's your value and your ability to protect you. And as you protect with intellectual property, you can protect your research or research. It's clear that you don't see it, but sometimes it's a simple way, it's brought in terms of communicative, and I've done it even today, the power of calculus. I've said it for impressioning you 606 petaflop, but nothing if I can use it. The key is the power of the software, the power of writing, and if we want to use that power to achieve a result. But if this is a distinctive factor, I can't think that this can acquire it, otherwise what I acquire is generally a commodity that I don't want to differentiate. And this is if you don't develop internally you can't activate that virtual circle that is I have the ability to eat, I develop, I have new challenges, I stimulate the resources, the ability to research and to solve problems that others don't solve. A this point, for me, we have seen where you started, what is the reality of today and how it is so important to work on these aspects of technology technology. Let's go ahead. We talk a lot about computer quantum mechanics, and to explain what is and what is. We will ask you a very general description, but what I want to ask you is what role will the development of this technology in the future, as regards the tasks that we offer to the supercalculate? Look, we are investing in the quantization of the computer. There are two fundamental reasons. I think that the quantization of the computer will allow us to solve things that are not now, at least not solveable, with the current architecture of supercalculation in times useful for the necessities that you have, because it is a big acceleration. Then the quantization of the computer will be much more, if we want, to our logic of of being human because we have more than a logical logic, just because we are close to atoms. We are made of atoms and therefore the polarization that we have is very close to our brain. Today all the calculations are proxy, are, if we want, we can do the simulation that we can do in the portal, but the logic is this. Then we have two important things. One, I said the challenge, What we debate now is the issue of the cryptographic keys. That's where we can apply it, but we can apply it in many ways, in the molecular parts, in the climatology, in the research, in the optimisation. Then you can see that the quantistic computer is something that we can play as competition, because they are just starting to find solutions that maybe didn't have a very profitable break. In the computer question there are two characteristics. One is that is small, based on the actual atoms, and the other is that requires little energy, which is what, in the supercalculation, requires much energy. All right, if we turn to the sustainability, you can understand that for some sectors, for some problems, the computer quantistic is ecstatic. It is not a general part, it is not a general part, it is not a general part, it is not a general part. And what we see? In the story that we have made to be technologically aic, even in the approach that we have on the energy level, and also on the supercalculation, we have an architecture hybrid, so there are some things that could continue to do in the evolution of the HPC, with the computer quantistic that can solve the more efficient and efficient problems that are not resolved. So, like we say, a mix energetico, but also a mix technological? Absolutely, absolutely. I thank you Dario, thank you Dario Pagani, Head of Digital Information Technology of ENI for this journey, this, let's say, this story of what is happening in the world of supercalculating, digital infrastructure. This episode of The Big Interview will end here, but will continue with the next episodes on all the Wired channels. Thank you. Thank you Emiliano.
