Supercomputers/Systems/Beaufix
France · upgraded · hpc
Beaufix
Also known as Meteo-France Beaufix; Beaufix2
Operated by Meteo-France at Meteo-France Computing Centre (Toulouse) .
Phase 1 seed dataset, compiled by hand. These rows were built from public operator, laboratory and vendor sources. A mechanical second-reader pass has since fetched every cited source: 1263 of 1345 systems have a readable citation that names them, and 56 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
- 2.16 PFlop/s
- Rpeak
- 2.59 PFlop/s
- Rmax ÷ Rpeak
- 83.5%
- Cores
- 73,440
- Power
- 830.2 kW
- Per watt
- 2.6 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 | - Bull bullx DLC cluster, 1,008 compute nodes | Bull | - | Reported | metis.upmc.frtop500.org |
| CPU | Intel Xeon E5-2697 v2 CPUs x86-64 · 12C · 2.7 GHz · 22 nm · 130 W Intel Xeon E5-2697v2 12C 2.7GHz | Intel | - | Reported | top500.orgscientific-computing.com |
| Interconnect | Mellanox InfiniBand FDR Interconnect Fat tree · 56 Gb/s per 4x port · About 1 microsecond Infiniband FDR | Mellanox NVDANASDAQ
rolls up to NVIDIA | - | Reported | top500.org |
| Cooling | - Bull direct liquid cooling of the bullx DLC blades | Bull | - | Reported | scientific-computing.com |
The read
The research and development half of the first scalar supercomputer pair that Meteo-France bought from Bull, a bullx DLC cluster at the Meteopole in Toulouse that replaced the agency NEC vector machines. Meteo-France staff describe Beaufix as a Bull cluster of 1,008 compute nodes in 72-node racks, the development and research machine, with Prolix as its operational twin at another site; Scientific Computing World reported the first bullx DLC system installed in Toulouse with a 500 TFlop/s Linpack result on Intel Xeon E5-2600 v2 processors with direct liquid cooling, and Clubic reported the November 2012 order of about 1 PFlop/s for 2013 to 2014 rising above 5 PFlop/s by 2016. TOP500 lists the 2013 machine as Beaufix (Bullx DLC B710, Xeon E5-2697v2 12-core 2.7 GHz, InfiniBand FDR, 24,192 cores, 469 TFlop/s, 391 kW) from the November 2013 list at rank 62, and a 2016 expansion as Beaufix2 (bullx DLC 720, Xeon E5-2698v4 20-core 2.2 GHz, 73,440 cores, Rmax 2.16 PFlop/s, 830 kW, TOP500 system 178962), last listed in November 2020; the figures recorded are those of the expanded Beaufix2. The Taranis and Belenos pair (already in this dataset) was installed in 2020. The two independent sources (Meteo-France and TOP500) agree on the vendor, name, site and size.
Timeline
- 2013 First operational source
Site & facility
Meteo-France Computing Centre (Toulouse)
- Location
- Toulouse, France
- Commissioned
- 2020
Change history
- 2026-10-09 Added Beaufix
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 |
|---|---|---|
| top500.org | names system and part | Intel Xeon E5-2697 v2, Mellanox InfiniBand FDR |
| metis.upmc.fr | does not name this system | - |
Checked 2026-10-13 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.