Supercomputers/Systems/Lichtenberg II Phase 2
Germany · operational · hpc
Lichtenberg II Phase 2
Also known as Lichtenberg II Cluster Stage 2
Operated by TU Darmstadt at TU Darmstadt Lichtwiese Campus .
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 1581 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
- 3.10 PFlop/s
- Rpeak
- 4.03 PFlop/s
- Rmax ÷ Rpeak
- 76.9%
- Cores
- 59,904
- Power
- -
- Per watt
- - 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.
How it compares
Ranked against the other machines in this dataset that report the same figure: those that first came into service within two years of Lichtenberg II Phase 2, and those in Germany. Computed from our tables; a metric is ranked only when at least eight peers report it.
| Metric | Value | Against its era | Against its country |
|---|---|---|---|
| Rmax | 3.10 PFlop/s | #220 of 257 · bottom 15% | #37 of 95 · top 38% |
| Rmax ÷ Rpeak | 76.9 % | #102 of 256 · top 40% | #38 of 94 · top 40% |
Every metric is ranked largest-first, so a high power-draw rank means a big power draw, not a good one. Rmax ÷ Rpeak uses HPL-measured results only. Rankings move as machines are added; they describe this dataset, not the world.
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 | - Lenovo ThinkSystem SD650 V3 (TOP500 manufacturer) | Lenovo | - | Reported | top500.org |
| Accelerator | Intel Data Center GPU Max 1550 Accelerators oneAPI (Xe-HPC) · Intel 7 / TSMC N5 · 600 W Intel Data Center GPU Max 1550 (Ponte Vecchio) accelerator nodes | Intel | - | Reported | hrz.tu-darmstadt.de |
| Accelerator | NVIDIA H100 SXM5 Accelerators CUDA (Hopper GH100) · 16,896C · 1.98 GHz · TSMC 4N · 700 W NVIDIA H100 SXM5 80GB accelerator nodes | NVIDIA | - | Reported | hrz.tu-darmstadt.de |
| Accelerator | AMD Instinct MI300X Accelerators ROCm (CDNA 3) · TSMC N5/N6 · 750 W AMD Instinct MI300X 192GB accelerator node | AMD | - | Reported | hrz.tu-darmstadt.de |
| CPU | - Intel Xeon Platinum 8470Q 52C 2.1GHz, two per node, 576 MPI nodes | Intel | 1,152 reported | Verified | hrz.tu-darmstadt.detop500.org |
| Interconnect | NVIDIA InfiniBand HDR Interconnect Fat tree or dragonfly+ · 200 Gb/s per port · About 0.6 microseconds InfiniBand HDR100 (MPI section) | NVIDIA | - | Verified | hrz.tu-darmstadt.detop500.org |
| Operating system | Red Hat Enterprise Linux Operating systems Linux distribution · Open source, commercial support subscription RHEL 8.6 | Red Hat | - | Reported | top500.org |
The read
The second stage of TU Darmstadt's Lichtenberg II (the first stage is lichtenberg-ii-phase-1, a Cascade Lake MEGWARE cluster). The university computing centre's hardware page lists a 576-node MPI section of two Xeon Platinum 8470Q (52 cores each, 512 GB DDR5-4800) on InfiniBand HDR100, three large-memory nodes of up to 6 TB, and a 9-node accelerator section with Intel Data Center GPU Max 1550 (Ponte Vecchio), NVIDIA H100 and AMD MI300X nodes. 576 nodes x 104 cores is exactly the 59,904 cores in TOP500's entry (system 180136, ThinkSystem SD650 V3, Lenovo, RHEL 8.6, 3.10 PFlop/s Rmax, 4.03 PFlop/s Rpeak), so TOP500 covers the MPI section only and the accelerator nodes are not in the benchmark figures. TOP500 first ranked it 230th in June 2023 and has it at 382nd in June 2025 and 409th in November 2025; the computing centre says the MPI section has been in testing since 5 December 2023, which post-dates that first ranking, so only the year 2023 is recorded as the start. No power figure is listed by TOP500.
Timeline
- 2023 First operational source
Site & facility
TU Darmstadt Lichtwiese Campus
- Location
- Darmstadt, Germany
- Cooling
- Hot-water cooling with waste heat to district heating
- Commissioned
- 2020
Change history
- 2026-10-29 Added Lichtenberg II Phase 2
Nearby in scale
The machines closest to Lichtenberg II Phase 2 by Rmax, smaller to larger.
- ASEAN-Korea HPC Facility · Indonesia, operational · 3.1 PFlop/s
- Komondor · Hungary, operational · 3.1 PFlop/s
- Medusa (Dell) · USA, operational · 3.08 PFlop/s
- Matrix · USA, operational · 3.11 PFlop/s
- IronMan (Dell) · USA, operational · 3.11 PFlop/s
- Thunder (AFRL) · USA, operational · 3.13 PFlop/s
Related systems
Same site
- Lichtenberg II · Germany, operational, 2020
Same operator
- Lichtenberg I Phase 2 · Germany, operational, 2014
- Lichtenberg I · Germany, decommissioned, 2013
Same accelerator
- IronMan (Dell) · USA, operational, 2024
- Vultr AMD GPU Supercompute Cluster (Chicago) · USA, operational, 2024
- MAXIMUS-384 · USA, operational, 2025
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.