Supercomputers/Systems/JEDI
Germany · operational · mixed
JEDI
Deep dive: JUPITER: Europe's exascale, American silicon
Also known as JUPITER Exascale Development Instrument
Operated by Forschungszentrum Julich at Julich Supercomputing Centre .
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: 1092 of 1153 systems have a readable citation that names them, and 40 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
- 4.50 PFlop/s
- Rpeak
- 5.13 PFlop/s
- Rmax ÷ Rpeak
- 87.7%
- Cores
- 19,584
- Power
- 67.31 kW
- Per watt
- 66.9 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 | - Eviden BullSequana XH3000 | Eviden | - | Reported | fz-juelich.detop500.org |
| Accelerator | NVIDIA GH200 Grace Hopper Superchip Accelerators Arm CPU + CUDA GPU · TSMC 4N NVIDIA GH200 Grace Hopper Superchip | NVIDIA | - | Reported | fz-juelich.detop500.org |
| CPU | NVIDIA Grace CPUs Armv9 (Neoverse V2) · 72C · 3.1 GHz · TSMC 4N Grace Hopper Superchip 72C 3GHz | NVIDIA | - | Reported | top500.orgfz-juelich.de |
| Interconnect | NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds Quad-Rail NVIDIA InfiniBand NDR200 | NVIDIA | - | Reported | top500.org |
| 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 (hot water) cooling with heat reuse | Eviden | - | Reported | fz-juelich.de |
The read
The single-rack precursor to the JUPITER booster module, built to the same BullSequana XH3000 and NVIDIA GH200 recipe so software and operations could be validated before the full exascale machine arrived. It topped the Green500 list twice in a row, June and November 2024, at essentially the same efficiency, about 72.7 GFlops/Watt, on only 67 kW of power, a small enough draw that FZJ's own account frames it in kilowatts rather than megawatts, unusual company for a system that shares its architecture with an exascale machine. TOP500 and FZJ's own press release independently confirm the same Rmax, Rpeak, core count and power figures, so the tier is verified.
Timeline
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Site & facility
Julich Supercomputing Centre
- Location
- Julich, Germany
- Cooling
- Direct liquid cooling
- Commissioned
- 2004
Change history
- 2026-09-24 JEDI added (JUPITER Exascale Development Instrument, FZJ, Germany)
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 | NVIDIA GH200 Grace Hopper Superchip, NVIDIA Grace, NVIDIA InfiniBand NDR (Quantum-2) |
| fz-juelich.de | names system and part | NVIDIA GH200 Grace Hopper Superchip, NVIDIA Grace, Direct liquid cooling, warm water, NVIDIA InfiniBand NDR (Quantum-2) |
Checked 2026-10-09 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.