Supercomputers/Systems/LiDO4
Germany · operational · mixed
LiDO4
Also known as LiDO4 HPC cluster
Operated by TU Dortmund University at TU Dortmund Campus Data Center .
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
- 3.14 PFlop/s
- Rpeak
- 4.56 PFlop/s
- Rmax ÷ Rpeak
- 68.9%
- Cores
- -
- Power
- 120 kW
- Per watt
- 26.2 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 | - NEC (TOP500 manufacturer NEC/Lenovo) | NEC | - | Reported | top500.org |
| Integrator | - Lenovo (TOP500 manufacturer NEC/Lenovo) | Lenovo | - | Reported | top500.org |
| Accelerator | NVIDIA H200 SXM Accelerators CUDA (Hopper GH100) · 16,896C · 1.98 GHz · TSMC 4N · 700 W 68 NVIDIA H200 GPUs with 141 GB each in 17 GPU nodes | NVIDIA | 68 reported | Reported | veranstaltungen.tu-dortmund.deitmc.tu-dortmund.de |
| CPU | - AMD EPYC 9334 32C 2.7 GHz on the GPU partition | AMD | - | Reported | top500.org |
| Interconnect | NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds InfiniBand NDR400 | NVIDIA | - | Reported | top500.org |
| Operating system | Rocky Linux Operating systems RHEL-compatible Linux, community-maintained · Downstream rebuild of Red Hat Enterprise Linux source Rocky Linux 9 | Rocky Enterprise Software Foundation | - | Reported | top500.org |
The read
TU Dortmund's fourth-generation LiDO Linux HPC cluster, replacing LiDO3 and described by the university as nearly nine times faster. The university's technical-data page lists 110 CPU-only nodes and 17 GPU nodes with 68 NVIDIA H200 GPUs (141 GB each), 7,680 compute cores, 74 TB CPU and 10 TB GPU memory, 2.7 PB of disk, 3.5 PFlop/s peak and 120 kW; funding is 4.5 million euros from the DFG and the State of North Rhine-Westphalia plus 900,000 euros from the university, with hot-water cooling at about 40 C. The ITMC ranking article says it was inaugurated in March 2026 yet that commissioning is due in the second half of 2026 with installation in its final stages, so first_operational_date is left blank. It ranked 417th on the June 2026 TOP500 (list full for our table, so no list row) as 'LiDO4-GPU Partition' (NEC/Lenovo, AMD EPYC 9334 32C, InfiniBand NDR400, Rocky Linux 9, 1,088 cores for the 17 GPU nodes) with 3.14 PFlop/s Rmax (ITMC quotes 3.145), 4.56 PFlop/s Rpeak and 59.8 kW for the run, plus HPCG 46 TFlop/s (126th) and 47th on the Green500 at 52.6 GFlops/W. The TOP500 Rpeak of 4.56 PFlop/s is higher than the university's 3.5 PFlop/s system peak and neither source explains why; the 59.8 kW run draw likewise differs from the 120 kW system figure. TOP500 entry: https://www.top500.org/system/180482/.
Timeline
- 2026-03 Announced source
Site & facility
Change history
- 2026-10-17 Added LiDO4
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 |
|---|---|---|
| itmc.tu-dortmund.de | names system, part and count | NVIDIA H200 SXM · counts: 68 |
| veranstaltungen.tu-dortmund.de | blocked our fetch | - |
Checked 2026-10-18 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
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.