Supercomputers/Systems/SuperMUC-NG Phase 2
Germany · operational · mixed
SuperMUC-NG Phase 2
Also known as SuperMUC-NG PH2
Operated by LRZ at Leibniz Supercomputing Centre, Garching .
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
- 17.19 PFlop/s
- Rpeak
- 27.96 PFlop/s
- Rmax ÷ Rpeak
- 61.5%
- Cores
- 149,760
- Power
- 725.9 kW
- Per watt
- 23.7 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 | - 240 Lenovo ThinkSystem SD650-I V3 Neptune DWC nodes | Lenovo | 240 reported | Verified | doku.lrz.detop500.org |
| Accelerator | Intel Data Center GPU Max 1550 Accelerators oneAPI (Xe-HPC) · Intel 7 / TSMC N5 · 600 W 960 Intel Data Center GPU Max 1550 (Ponte Vecchio), four per node | Intel | 960 reported | Reported | doku.lrz.de |
| CPU | - two Xeon Platinum 8480+ per node, 240 nodes | Intel | 480 estimated | Reported | doku.lrz.de |
| Interconnect | NVIDIA InfiniBand HDR Interconnect Fat tree or dragonfly+ · 200 Gb/s per port · About 0.6 microseconds NVIDIA/Mellanox HDR InfiniBand 200 Gb/s fat tree | NVIDIA | - | Reported | doku.lrz.de |
| Storage | Intel DAOS Storage Object store, no POSIX layer underneath · 230 PB on Aurora · About 31 TB/s 1 PB DAOS tier at 750 GB/s alongside 50 PB and 20 PB file systems | Intel | - | Reported | doku.lrz.de |
| Cooling | Direct liquid cooling, warm water Cooling Direct-to-chip cold plates, facility water loop · Removes 90+ percent of node heat to liquid Lenovo Neptune direct warm-water cooling | Lenovo | - | Reported | doku.lrz.de |
The read
The GPU extension of LRZ's flagship: 240 Lenovo ThinkSystem SD650-I V3 nodes, each with two Xeon Platinum 8480+ and four Intel Data Center GPU Max 1550 (960 GPUs, 128 GB HBM2e each), on HDR InfiniBand with a 1 PB DAOS tier alongside 50 PB and 20 PB file systems, warm-water cooled. First ranked 51st on the November 2023 TOP500 and 127th on the June 2026 list. LRZ's documentation gives the hardware and TOP500 the 17.19 PFlop/s Rmax, 27.96 PFlop/s peak, 149,760 cores and 725.9 kW; TOP500 alone supplies those performance and core figures, so the tier is reported. LRZ's news page described the machine as in pilot operation in early 2024, so no first-operational date is recorded.
Timeline
- 2023-11 List debut at #51 (November 2023) 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
Leibniz Supercomputing Centre, Garching
- Location
- Garching, Germany
- Cooling
- Direct warm-water cooling
- Commissioned
- 2006
Change history
- 2026-10-02 Added SuperMUC-NG Phase 2 at LRZ
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 |
|---|---|---|
| doku.lrz.de | blocked our fetch | - |
| top500.org | names system and part | Intel Data Center GPU Max 1550, NVIDIA InfiniBand HDR |
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.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.