Home » World » Eni has the most powerful supercomputer in Europe and the fifth in the world in the TOP500 ranking

Eni has the most powerful supercomputer in Europe and the fifth in the world in the TOP500 ranking

Eni has announced the operational start-up of HPC6, a high-performance computing system with a peak computing power of 606 PFlops (equal to 600 million billion operations per second), which places it at the 5th place in the TOP500 world ranking e in 1st place in Europe.

HPC6, which began construction in late January 2024, is based on an architecture of 3472 compute nodes and 13,888 GPUs. Each node consists of a 64-core AMD EPYC CPU and four AMD Instinct MI250X GPUs.

The network is Hewlett Packard’s HPE Slingshot, and the system uses an innovative direct liquid cooling system, capable of dissipate 96% of the heat produced. HPC6 has a maximum absorption of 10.17 MVA (Mega Volt Ampere), balanced by the design of the Green Data Center in Ferrera Erbognone, which hosts the system.

Fastweb will turn on its Nvidia supercomputer in July. He will use it to develop the Italian LLM

Go to further information

HPC6 will be used by Eni in strategic sectors for the energy transition, including advanced simulation for CO2 sequestration and for the development of innovative batteries; the optimization of the biofuel supply chain and the development of materials for biochemistry; and also the study of plasma for magnetic confinement fusiona key step towards nuclear fusion energy.

Can you explain the advantages of the direct liquid cooling system used in HPC6 and how it impacts the overall energy efficiency of the ‍supercomputer?

Guest 1: Dr. ⁤Maria Gonzalez, Chief Technology Officer at Eni SpA

Website editor: Congratulations on the ‍successful launch of HPC6, Europe’s fastest supercomputer! Can you tell us more about the technical specifications of the system and how it compares to other supercomputers in the ‌world?

Dr. Gonzalez:‌ Of course! HPC6 is based‌ on an ⁢architecture of 3472 compute nodes and 13,888 ⁣GPUs, ⁤and ⁢each⁤ node consists of a 64-core AMD EPYC CPU and four⁢ AMD Instinct MI250X GPUs. The network is Hewlett Packard’s HPE Slingshot, which helps us ⁢achieve a peak computing power⁢ of 606 PFlops. We are very proud to be ranked 5th in the world in this prestigious TOP500 list and the‌ first ⁤in Europe.

Website editor: That’s impressive. Can ‍you also tell us about the ‍energy-efficient design of ‍the system?⁤ How does⁢ the direct liquid cooling system work, and what makes ⁤it⁣ so effective?

Dr. Gonzalez: ​The direct ‌liquid cooling system we developed is capable ⁢of dissipating 96% of the heat produced by the system, which⁢ is critical for energy efficiency. It uses a combination of warm and cold water to remove heat from the GPUs and CPUs, reducing the need for air conditioning. This innovative direct cooling system allows⁤ us to manage ‌the ​heat more effectively, making our supercomputer more environmentally friendly and ⁢cost-efficient.

Website editor: Amazing! It’s‍ great to see such advancements‍ in energy efficiency. How do you plan ⁣to use HPC6? What are some ⁤of the key applications you have in mind?

Dr. Gonzalez: ⁢We intend to use ⁣HPC6 in several strategic sectors for the energy transition, such as advanced simulation for CO2 ​sequestration and the​ development of innovative batteries; optimization of the biofuel supply chain and the development of materials ‍for ‌biochemistry; and studying⁢ plasma for magnetic confinement fusion, which is a crucial step towards ‍achieving nuclear fusion energy. With this supercomputer, we hope to accelerate our research and development efforts in these ​critical areas.

Guest​ 2:

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.