The new national supercomputer, DAIDALOS, is now operational, and its official inauguration will take place on Monday, October 19. Housed in the new Data Center at the Lavrio Technology and Culture Park of the National Technical University of Athens, it will provide computing resources to universities, research centers, startups, small and medium-sized enterprises, and public agencies.
The project is an initiative of the Ministry of Digital Governance and Artificial Intelligence and is being implemented by the National Network for Technology and Research Infrastructure (EDYTE S.A. – GRNET). The total budget amounts to 58.9 million euros. The project is part of the National Recovery and Resilience Plan “Greece 2.0” and is funded with the support of the European Union through NextGenerationEU and the EuroHPC Joint Undertaking (JU).
ΣΧΕΤΙΚΑ ΑΡΘΡΑ
DAIDALOS’s measured performance reaches 85.69 petaflops, ranking it 31st globally on the TOP500 list and 23rd on the Green500 list in terms of energy efficiency. At the same time, it serves as the core of the Greek PHAROS Artificial Intelligence facility, which links the computing infrastructure with the needs of Artificial Intelligence application development.
“Daedalus is being plugged in. Essentially, the final adjustments are now underway, and it is being put into operation. Daedalus is a machine capable of training models. There are many universities, research institutions, and companies that have an idea in mind or have tested something on a smaller scale but now need space, time, and, of course, at no cost, to train their models. So this will function as a beacon. It will call out more loudly and more openly to the entire artificial intelligence ecosystem to build models and experiment with them. What comes next is where these models will be deployed. Because here, too, the economic issue comes into play with great urgency. “Artificial intelligence isn’t cheap,” Dimitris Papastergiou, Minister of Digital Governance and Artificial Intelligence, told the Athens-Macedonian News Agency (ANA-MPA).
He continued: “In October, the bidding process will begin for the second machine in Kozani, which will be an equally powerful machine, but primarily designed to run these models. Until then, of course, the models can run on the infrastructure provided by Pharos, but we’re bringing in an equally powerful machine because the demands will multiply. The models that will be trained over the next few years will eventually need to be run at high performance levels. So I think that will be the time when we’ll be able to put the second machine into operation as well. At the same time, of course, there are other machines. For example, there’s Ikaros, which isn’t very well known and is the new machine from the National Technical University of Athens. “So this, too, becomes part of the broader ecosystem and falls under the Pharos umbrella so that, together with Daedalus, we can provide services to this entire community.”
How a Supercomputer Works
A supercomputer consists of many powerful computing units that operate simultaneously and work together as a single system. This allows it to perform complex calculations and process large volumes of data in much less time than a conventional computer.
Depending on the application, a task that would take weeks or months can be completed in hours or days. This capability is particularly important for training large artificial intelligence models, scientific simulations, and large-scale data analysis, as well as for applications that require the examination of many different scenarios.
Until now, Greek research teams, universities, and companies requiring significant computing power often sought access to infrastructure abroad. The launch of DAIDALOS expands the capabilities for carrying out such tasks within the country, enabling the development of research projects and applications that require significant computational resources.
Applications in civil protection, health, and industry
In the fields of civil protection and climate change research, DAIDALOS can support high-resolution simulations of natural disasters. A research team can, for example, examine the likely progression of a wildfire under different weather conditions by varying parameters such as wind speed. Computing power allows for the examination of more scenarios in less time and with greater accuracy.
In healthcare, a Greek startup can use the infrastructure to train artificial intelligence models that analyze medical images, such as X-rays and CT scans. Developing such applications requires processing large volumes of data and repeated cycles of testing and evaluation—processes that can be accelerated through the use of powerful computing systems.
Similarly, a university laboratory can process large collections of Greek texts and train language models that better account for the unique characteristics of the Greek language. In this way, applications tailored to the country’s needs can be developed, without being limited to general-purpose solutions designed for languages with a much larger volume of available data.
In industry, computing power can be leveraged to create digital twins of products and production processes. Companies can simulate different conditions before proceeding with actual manufacturing, identify potential problems earlier, and improve the design, thereby reducing time and costs.
Applications also include biomedical research, physics, chemistry, the development of new materials, climate modeling, large-scale data analysis, and the development of artificial intelligence applications for the public and private sectors.
DAIDALOS and PHAROS: Computing Power and Its Utilization
DAIDALOS serves as the computational core of the Greek AI Factory PHAROS. The two infrastructures have distinct but complementary roles: the supercomputer provides the required computing power, while PHAROS organizes access to it and combines it with data, tools, expertise, technical support, and partnerships.
The goal is to enable businesses, universities, and research teams to leverage available resources for the development and evaluation of artificial intelligence applications. Linking computing power with data and scientific knowledge facilitates the use of this infrastructure by the research and innovation ecosystem.
PHAROS places Greece among the 14 countries worldwide that possess infrastructure of this scale for the development of next-generation artificial intelligence applications.
The development of such infrastructure is in line with the direction set by the Mitsotakis government to strengthen the country’s capabilities in artificial intelligence. The goal is for Greece not to be limited to using technologies developed abroad, but to have the infrastructure and human resources to create new applications.
In the historic building of the former Power Station
The new Data Center is housed in the historic building of the former Power Station, located in the Lavrio Technology and Cultural Park of the National Technical University of Athens (NTUA). The space, which previously housed industrial infrastructure, is now home to DAIDALOS, designed to support the modern needs of research and technology.
The infrastructure can be utilized by universities and research centers for complex scientific simulations, by startups for training artificial intelligence models, and by companies for developing and testing new products and services.
As the leadership of the Ministry of Digital Governance and Innovation emphasizes at every opportunity, with DAIDALOS, Greece is acquiring large-scale computing resources within its borders, which research and business entities can utilize for projects that, until now, often required access to similar infrastructure abroad.
Source: APE-MPE
