The power of HPC5, to help fight the pandemic

The great advantage of technological development is the many incredibly useful tools it gives mankind, which can forge unexpected connections, above all at times of crisis. Connections like that between Eni and the European Union, for example, who have come together to fight Covid-19. Fight it how, you ask? By putting the funds aside to form a network of researchers, universities and pharmaceutical companies of the highest calibre, and putting one of the world’s most powerful calculators to work.

These are the pillars of Exscalate4CoV, a public-private cooperative project funded by the European Union’s “Horizon 2020” initiative, to fight the coronavirus pandemic and improve the management and care of its patients. The E4C consortium, overseen by the Italian pharmaceutical company Dompé Farmaceutici, assembles 18 institutions and research centres in seven European countries and uses the supercomputer HPC5 from Eni’s Green Data Centre. The name of the project is an acronym: “Exscalate” is short for “EXaSCale smArt pLatform Against paThogEns”, one of the most powerful and energy-efficient supercomputing platforms in the world. E4C has a chemical library of 500 billion molecules and a processing capacity of over 3 million molecules a second. It provides an essential resource when searching for a vaccine against a pathogen.

Eni’s involvement in a project of this scale is thanks wholly to the scientific research and technological innovation it has always made a priority. In 2020, in fact, the Italian energy company revealed its supercomputer HPC5 (short for high-performance computing, layer 5). It’s a super-brain, the cutting edge, one of the 10 most powerful computers in the world and the top not in the hands of a government Combined with its predecessor HPC4, it can reach calculation speeds of 70 million billion operations a second (petaflops). That’s an impressive figure, especially when you consider that on its own, HPC4 runs at a third of the newcomer’s speed, taking two months to process the same amount of information the HPC5 does in 10 days.

This is a giant leap and a springboard to the theoretical chemistry research under way to find a vaccine against Covid-19. Mapping the molecule of a virus and its interactions with molecules entails millions of pairings and checks, and painstaking gathering and cataloguing of data; only a system of processors as powerful as HPC5 can do it in a reasonable amount of time. But another great advantage of this machinery is the way it’s powered. When it comes to energy saving, HPC5 is the finest computer the world has ever seen. Given that it can do almost 20 million operations a second on just one watt of electricity, it’s no exaggeration to call it futuristic. Eni built it to study the lay of the earth, using sophisticated algorithms that produce very detailed geophysical models. HPC5 has come out with models of the subsoil that are so accurate they can reveal things hidden many miles below the surface. That’s how Eni found Zohr, the biggest gas field ever discovered in the Mediterranean.

But despite being designed and so far used for a very different end, in our current predicament HPC5 has provided science with a useful ally. The fact it’s in Europe is another plus. Just a few weeks after the Covid-19 epidemic broke out, President of the European Commission Ursula von der Leyen identified one of the EU’s priorities as having an independent technological infrastructure. One way to that independence might be having proper machinery and software of its own, rather than resorting to technology from private companies or governments outside the bloc. It would also avoid the risks that the great traffic of online data now brings, and that’s where the consortium Exscalate4CoV’s strength lies. It’s attempting to bring different players in Europe together for a good cause, to fight our present epidemiological crisis. Francesco Frigerio, a researcher at Eni, explains: “Eni has signed up to the project Exscalate4CoV, providing its skills and experience in research and the computing power of its industrial supercomputer HPC5. The goal of Exscalate4CoV is to bring together skills and technology from around Europe to take this pandemic on quickly and as best we can, finding the safest and most promising drugs in the fight against Covid-19.”

But what kind of work will this mighty machinery be put to exactly? “We began working with Cineca, one of the best computing research centres in the world,” Frigerio goes on, “doing molecular dynamics simulations of the proteins on the surface of SARS-CoV2, which play a key role in the virus’s infection mechanism. The data we get from modelling these proteins’ behaviour on HPC5 will be used in the next stage of work, led by Dompé, in which we’ll compare their structures with those in a database containing the 10,000 active ingredients of pharmaceutical compounds known today. That’s how we plan to find which of these active ingredients, perhaps an active ingredient developed to treat a completely different disease, could interact with the virus and block it.”

In other words, HPC5 will simulate the dynamic behaviour of the molecules and study how they interact with each other (what’s known in the field as “molecular modelling”). We will then be able to see how, for example, the virus responds to the active ingredient in a given drug. These calculations demand an intimidating processing power; in fact “If we had to do this work without the help of a large supercomputer, it’d take us years, but we’ll probably manage to model about 30 of the proteins in Covid-19 in a couple of months, then hand things over to other researchers who will do experiments. We trust we’ll have found an effective drug by the end of 2020,” Frigerio goes on.

What will make the difference, according to researchers, is the software chosen to process the data. It’s called Gromax and it’s an open source (meaning unowned and free to edit or share) molecular dynamics simulation programme that can make full use of HPC5’s 7,280 GPU (graphics processing unit, optimised for calculating large amounts of data simultaneously), bringing added value to the research project. Not only are different companies and institutions joining the network, but a lot of care is being taken over the environmental impact of the research, which is being done along collaborative lines. Within the EU, government bodies, scientific companies, foundations and universities have come together, through the European Commission, to act as incubators and key financers. This is no small achievement when you consider all the fear provoked by the pandemic and the commercial interest behind the current search for a vaccine. The fact that these bodies from around Europe have come together to work actively for the common good should surely be a point of pride for the EU and the parties involved. Even the modus operandi chosen by the researchers encourages international cooperation. Not for nothing did they choose an open source model, which lets you exchange data and results and see others’ progress. Their approach is decisively collaborative and far from speculative.

“Personally I believe strongly in the ethical value of open source and open knowledge in general,” says Frigerio. “When it comes to medicine, as with vaccines for blocking coronavirus, we’d be nowhere without open science, which makes all the genome sequences of all the viruses linked to SARS-CoV2 available as research repositories containing the structure of the proteins. No one working on their own could have got the results we’re all reaching together. Besides financial support and HPC5, Eni has put its researchers’ expertise on the table. Now more than ever,” he concludes, “aside from financial resources, we need knowledge to be as open as possible, plus expertise that once combined will stop the spread of this new virus.”