Supercomputers/Systems/Leonardo
Italy · upgraded · mixed
Leonardo
Quantum: linked to Pasqal SOL (140 neutral atom qubits), IQM Radiance NOX at CINECA (54 superconducting qubits)
Operated by CINECA at Bologna Technopole .
Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1540 systems have a readable citation that names them, and 64 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.
Measured performance
- Rmax
- 241.2 PFlop/s
- Rpeak
- 306.3 PFlop/s
- Rmax ÷ Rpeak
- 78.7%
- Cores
- 1,824,768
- Power
- 7.49 MW
- Per watt
- 32.2 GF/W
Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.
What this machine is made of
One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.
| Role | Part | Supplier at build | Quantity | Confidence | Source |
|---|---|---|---|---|---|
| Integrator | - Atos BullSequana XH2000 | Eviden | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| Accelerator | NVIDIA A100 SXM4 64GB Accelerators CUDA (Ampere GA100) · 6,912C · 1.41 GHz · 7 nm · 400 W NVIDIA A100 SXM4 64 GB, 4 per booster node | NVIDIA | 13,824 reported | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| CPU | Intel Xeon Platinum 8358 CPUs x86-64 · 32C · 2.6 GHz · 10 nm · 250 W Xeon Platinum 8358 32C 2.6GHz | Intel | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| Interconnect | NVIDIA InfiniBand HDR Interconnect Fat tree or dragonfly+ · 200 Gb/s per port · About 0.6 microseconds NVIDIA HDR100 InfiniBand, dragonfly+ | NVIDIA | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| Storage | DDN EXAScaler Storage Lustre appliance DDN Lustre appliance | DDN | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| Cooling | Direct liquid cooling, warm water Cooling Direct-to-chip cold plates, facility water loop · Removes 90+ percent of node heat to liquid direct liquid cooling | Eviden | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| Power | Vertiv Liebert UPS Power Large-frame UPS, typically N+1 or 2N · Up to several MW per frame Vertiv provides power and cooling solutions for Leonardo | Vertiv | - | Reported | it-online.co.za |
| Operating system | Red Hat Enterprise Linux Operating systems Linux distribution · Open source, commercial support subscription Red Hat Enterprise Linux | Red Hat | - | Reported | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 13,824 accelerators | 2026-10 | Operator stated | Booster module: 3,456 nodes x 4 A100 SXM4 64 GB. | leonardo-supercomputer.cineca.eueurohpc-ju.europa.eu |
| IT power | 10 MW | 2026-10 | Operator stated | IT load of the Bologna Technopole data centre for stage 1 (2021-2025), shared with other CINECA systems; stage 2 (2025-2030) is stated as 20 MW. | leonardo-supercomputer.cineca.eu |
| PUE | 1.1 | 2026-10 | Design target | CINECA says the facility PUE is 'estimated below 1.1' (warm-water cooling, 8 MW liquid-cooling capacity). Estimate, not metered. | leonardo-supercomputer.cineca.eu |
The read
The European counterweight to LUMI, and a different supply chain: Eviden integration on the BullSequana XH2000 platform, Intel hosts, NVIDIA A100 accelerators in a custom 64 GB variant, NVIDIA HDR InfiniBand. Eviden is the corporate descendant of Bull, so this row is one of the places where the vendor-alias table earns its keep: an unresolved dataset would scatter the same supplier across Bull, Atos and Eviden.
Timeline
- 2019 Announced source
- 2022-11 List debut at #4 (November 2022) source
- 2022-11 First operational source
- 2024-06 Data-centric partition added source
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Site & facility
Bologna Technopole
- Location
- Bologna, Italy
- Cooling
- Direct liquid cooling
- Commissioned
- 2022
Change history
- 2026-10-04 Added AI datacenter metrics for leonardo
- 2026-09-22 Gained a Green500 rank pointer (#18 on the 2023-11 list)
Source check
We fetched this system's own citations and recorded whether each page actually mentions it. This is a corroboration signal, not a fact check, and it is published so you can see how well the sourcing holds up rather than take it on trust.
| Cited source | Result | Found on the page |
|---|---|---|
| eurohpc-ju.europa.eu | does not name this system | - |
| leonardo-supercomputer.cineca.eu | names system, part and count | NVIDIA A100 SXM4 64GB, Direct liquid cooling, warm water, Intel Xeon Platinum 8358, NVIDIA InfiniBand HDR · counts: 13824 |
Checked 2026-10-18 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
Power and cooling
Why air ran out at 20 kW a rack, how warm-water cooling removes the chillers, and what PUE conceals.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.