Galileo
Italian national Tier-1 supercomputer for scientific and technical computing
Galileo is a 1.1 petaFLOPS supercomputer at CINECA in Bologna, Italy, built by IBM on its NeXtScale architecture. Available from January 2015 and in full production from 2 February that year, it serves Italian scientific computing and European researchers through the PRACE infrastructure.
- 1.1 petaFLOPS IBM NeXtScale cluster at CINECA, Italy
- 516 compute nodes, 8,256 Intel Haswell cores in total
- 768 Intel Phi 7120p and 80 NVIDIA K80 accelerators
- Ranked 105th on TOP500 in June 2015
- Italy's national Tier-1 machine, also open to Europe via PRACE
What it is
Galileo is a hybrid supercomputer run by CINECA, an Italian computing consortium, at its site in Casalecchio di Reno near Bologna. Built by IBM using the NeXtScale platform, it delivers 1.1 petaFLOPS, meaning roughly 1.1 thousand trillion floating-point operations per second. The machine is Italy's national Tier-1 system for high-performance computing, a top-level national resource open to researchers. Its work covers computational fluid dynamics, materials and life sciences, and geophysics. Beyond Italy, scientists across Europe can use it as a Tier-1 system within PRACE, a partnership that shares major computing resources between countries. The cluster runs CentOS 7.0 and is reached through login nodes over ssh.
How it came about
The system became available at CINECA in January 2015 and entered full production on 2 February 2015. Funding came from Italy's Ministry of Education, Universities and Research, the Istituto Nazionale di Fisica Nucleare and the University of Milano-Bicocca. In June 2015 it placed 105th on the TOP500 ranking of the world's fastest machines, and by November 2015 it stood at 130th. On the Green500, which measures efficiency per watt, it reached 389th with a figure of 242.17 MFLOPS/W.
The compute
Galileo is an IBM Linux cluster with InfiniBand networking through 4x QDR switches. Its 516 compute nodes each hold two 8-core Intel Haswell processors running at 2.40 GHz, giving 8,256 cores overall, plus 128 GB of shared memory per node. The design is heterogeneous: 359 nodes carry Intel Phi 7120p accelerators, two each for 768 in total, while 40 nodes carry NVIDIA K80 GPUs, two each for 80 in total, of which 20 are set aside for scientific research. Eight login nodes, each with two NVIDIA K40 GPUs, handle access and visualisation, and eight NX360 M5 service nodes manage I/O and administration. Local scratch storage totals 2,000 TB.
Where the GPUs are
The machine sits at CINECA in Casalecchio di Reno, in the Bologna area of Italy. CINECA operates it on behalf of a consortium of members. Researchers connect remotely to the login nodes at login.galileo.cineca.it using ssh. Power draw for the system is recorded as 2,825.55 kW.
Customers and deals
Galileo serves Italian scientific and technical computing users as the national Tier-1 resource. It is also open to researchers from other European countries as a Tier-1 system within the PRACE infrastructure. Its stated application areas are computational fluid dynamics, materials and life science, and geophysics.
Money and ownership
The system was sponsored by Italy's Ministry of Education, Universities and Research, the Istituto Nazionale di Fisica Nucleare and the University of Milano-Bicocca. No cost figure is given.
Its place in AI
Galileo acts as Italy's national Tier-1 high-performance computing machine and feeds into European access through PRACE. Its peak speed of 1,103.1 PFLOPS placed it 105th on the TOP500 list in June 2015 and 130th by November 2015. Efficiency testing on the Green500 put it at 389th, at 242.17 MFLOPS per watt.
Questions and criticism
The system draws 2,825.55 kW of power, and its Green500 efficiency figure of 242.17 MFLOPS/W placed it 389th on that list.

Questions people ask about Galileo
What is Galileo?
It is a supercomputer at CINECA near Bologna, Italy, built by IBM on the NeXtScale platform. Rated at 1.1 petaFLOPS, it handles scientific and technical work such as fluid dynamics, materials and life science, and geophysics.
Who can use it?
It is Italy's national Tier-1 high-performance computing resource, open to Italian researchers. Scientists elsewhere in Europe can also access it as a Tier-1 system within the PRACE infrastructure.
What hardware does it use?
Each of its 516 compute nodes has two 8-core Intel Haswell processors and 128 GB of memory. Accelerators include 768 Intel Phi 7120p chips and 80 NVIDIA K80 GPUs, linked by InfiniBand with 4x QDR switches.
How did it rank?
In June 2015 Galileo reached 105th on the TOP500 list of the world's fastest computers, and 130th by November 2015. On the Green500 efficiency list it placed 389th at 242.17 MFLOPS per watt.
Good reading and sources
The facts on this page come from the Wikipedia article "Galileo (supercomputer)", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryNeocloud rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.
Like Galileo
All AI supercomputersMore from IBM
All of them
馃嚭馃嚫 United StatesSequoia
IBM Blue Gene/Q supercomputer built for US nuclear weapons simulation at LivermoreSupercomputer 路 Other and custom chips 路 2011