Germany · operational · ai training

DAIS

Also known as MPCDF Data and AI System

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: 6.03 PFlop/s
Rpeak: 9.22 PFlop/s
Rmax ÷ Rpeak: 65.4%
Cores: 19,584
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.

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 SR780a V3 Neptune	Lenovo 992HKEX	-	Reported	top500.org
Accelerator	NVIDIA H200 SXM Accelerators CUDA (Hopper GH100) · 16,896C · 1.98 GHz · TSMC 4N · 700 W NVIDIA H200 141 GB, 8 per node on 17 H200 nodes (derived from the node count)	NVIDIA NVDANASDAQ	136 reported	Reported	docs.mpcdf.mpg.de top500.org
Accelerator	NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W NVIDIA B200 180 GB, 8 per node on 16 B200 nodes (derived from the node count)	NVIDIA NVDANASDAQ	128 reported	Reported	docs.mpcdf.mpg.de
CPU	- Intel Xeon Platinum 8568Y+ 48C 2.3GHz	Intel INTCNASDAQ	-	Verified	docs.mpcdf.mpg.de top500.org
Interconnect	NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds Infiniband NDR, 8x200 Gb per GPU node	NVIDIA NVDANASDAQ	-	Reported	top500.org docs.mpcdf.mpg.de
Operating system	SUSE Linux Enterprise Server Operating systems Linux distribution · Open source, commercial support subscription SLES 15 SP6	SUSE	-	Reported	top500.org

The read

DAIS, the Max Planck Computing and Data Facility's GPU system for AI and data-intensive work for Max Planck institutes. MPCDF's user guide lists 33 GPU nodes: 17 with eight NVIDIA H200 (141 GB) GPUs, 16 with eight NVIDIA B200 (180 GB) GPUs and 2 with four RTX PRO 6000 GPUs, all on Intel Xeon Platinum 8568Y+ 48C hosts (the RTX PRO nodes use AMD EPYC 9555) with 8x200 Gb InfiniBand per GPU node, and the DAIS page says a first extension with Blackwell GPUs is underway. The TOP500 entry is the Lenovo ThinkSystem SR780a V3 Neptune H200 partition on InfiniBand NDR under SLES 15 SP6, with 19,584 cores, Rmax 6.03 PFlop/s and Rpeak 9.22 PFlop/s; it debuted at rank 247 on the November 2025 list and was 270th in June 2026. MPCDF and TOP500 agree on the 6.0 PFlop/s result, the rank and the H200 and Xeon 8568Y+ hardware, so the tier is verified. The H200 and B200 GPU counts are node counts multiplied by eight and are marked reported.

Timeline

2025-11 List debut at #247 (November 2025) source

Public list appearances

Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.

#247: TOP500, November 2025 →

Site & facility

MPCDF Garching

Location: Garching, Germany
Commissioned: 2021

mpcdf.mpg.de top500.org

Change history

2026-10-07 Added DAIS (mpcdf)

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 H200 SXM, NVIDIA InfiniBand NDR (Quantum-2)
mpcdf.mpg.de	names system only	-

Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.