Over the last few years, supercomputing or high performance computing (HPC) has been a strategic focus for the scientific and innovation ecosystem. Find out what this technology consists of and learn about its various applications.
The first supercomputer made available to the Portuguese scientific community was installed by FCCN, FCT's digital services, in 1988. Since then, over the last four decades, this area has been marked by great evolution and is one of the European Commission's priorities for investment in technology and innovation. The aim of this effort, which brings together public and private organizations , is to "develop a world-class supercomputing ecosystem in Europe", says the EuroHPC JU.
But what exactly is supercomputing or advanced computing?
As the name suggests, a supercomputer is a powerful machine, capable of solving some of today's most difficult problems in real time. For this reason, they are very relevant machines in today's context, allowing researchers to "find the answer to some of the most complex scientific questions", explains the European Commission.
The technology behind supercomputing is, of course, related to information processing, explains IBM on its website. Traditional computers use the "sequential computing" method, dividing up the work by performing a sequence of tasks on the same processor. In the case of a supercomputer, "massively parallel computing" is used, in which several tasks are carried out simultaneously, using tens of thousands of processors or cores.
For this reason, from the point of view of its architecture, a supercomputer uses clusters of networked nodes, with each node performing part of the same task. On its website, Amazon Web Services explains that a supercomputing cluster "consists of hundreds or even thousands of computing nodes", with each node "containing between 8 and 128 CPUs [Central Processing Units]".
The end result of this infrastructure is impressive - the world's most powerful supercomputers are capable of performing at least one hundred million billion (10¹⁷) operations per second. And there's a technological leap expected soon. The installation of JUPITER, Europe' s first exascale supercomputer, capable of performing 10¹⁸ operations per second, is scheduled for the second half of 2024.
This information processing capacity is highly relevant in a wide variety of areas. As the European Commission points out, it can be used to "advance the frontiers of knowledge in virtually every scientific field". From "controlling and mitigating the effects of the climate crisis" to "producing safer and more environmentally friendly vehicles", to "increasing cybersecurity", there are many applications for these machines in the world of science and innovation.
Another field in which supercomputing is already an essential tool is medicine. High-performance computing is used in the manufacture of new drugs, testing new candidate molecules or "fine-tuning" existing drugs. An often-shared example illustrates the power of these machines. The first attempt to sequence the human genome took 13 years. Today, a supercomputer can do it in less than a day.
Supercomputing in Portugal
Portugal stands out on the supercomputing scene with Deucalion, the fastest national supercomputer ever supported by the Recovery and Resilience Plan (PRR). In addition, Portugal 's participation in the Spanish supercomputer MareNostrum 5, funded by the RRP, reinforces international collaboration in this strategic area.
For more details on access, information can be found in the open euroHPC call, with a deadline for applications of September 6. As well as the MareNostrum 5 expression of interest, whose deadline is September 9.
Stay tuned for the opening of the 5th National Call for Advanced Computing Projects, which will take place from September 10th. To access these technologies: more information on the FCT calls page.
