Supercomputers/Systems/Viper-CPU
Germany · operational · hpc
Viper-CPU
Operated by MPCDF at MPCDF Garching .
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
- 4.29 PFlop/s
- Rpeak
- 4.80 PFlop/s
- Rmax ÷ Rpeak
- 89.4%
- Cores
- 97,920
- Power
- 784 kW
- Per watt
- 5.5 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 Viper-CPU, 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 | 4.29 PFlop/s | #182 of 234 · bottom 23% | #30 of 95 · top 31% |
| Energy efficiency | 5.5 GF/W | #113 of 121 · bottom 8% | #28 of 68 · top 40% |
| Rmax ÷ Rpeak | 89.4 % | #22 of 231 · top 9% | #10 of 94 · top 10% |
| Power draw | 784 kW | #60 of 126 · top 47% | #23 of 69 · top 32% |
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 | - Bull (TOP500 manufacturer) | Eviden | - | Reported | top500.org |
| CPU | AMD EPYC 9004 (Genoa) CPUs x86-64 · 96C · 2.4 GHz · TSMC N5 · 360 W 1,536 AMD EPYC 9554 64C 3.1GHz (two per node, 768 nodes) | AMD | 1,536 reported | Verified | docs.mpcdf.mpg.detop500.org |
| Interconnect | NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds Mellanox InfiniBand NDR200, non-blocking fat tree | NVIDIA | - | Verified | docs.mpcdf.mpg.detop500.org |
| Operating system | Red Hat Enterprise Linux Operating systems Linux distribution · Open source, commercial support subscription Red Hat Enterprise Linux | Red Hat | - | Reported | top500.org |
The read
The CPU-only cluster of the Max Planck Society's Viper at MPCDF in Garching, run as a logically separate cluster from Viper-GPU (see viper-gpu) with its own file systems. MPCDF's user guide lists 768 nodes with two AMD EPYC 9554 (Genoa, 3.1 GHz, 128 cores per node, 98,304 cores in total, 1,536 CPUs), about 432 TB of DDR5 in mixed 512 GB, 768 GB, 1 TB and 2.3 TB nodes, 4.9 PFlop/s theoretical FP64 peak, a non-blocking NDR200 InfiniBand fat tree, about 12 PB of disk and operation since June 2024. TOP500 (system 180413) lists Bull hardware with 97,920 cores (exactly 765 nodes of 128 cores), 4.29 PFlop/s Rmax, 4.80 PFlop/s Rpeak and 784 kW at measurement level 1, under the site name Max-Planck-Gesellschaft MPI/IPP; it ranks 299th in November 2025 and 328th in June 2026. TOP500 gives installation year 2025, whereas MPCDF says the CPU part has run since June 2024, which is the date used here. The integrator is taken from TOP500 alone, as for viper-gpu.
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
Change history
- 2026-10-29 Added Viper-CPU
Nearby in scale
The machines closest to Viper-CPU by Rmax, smaller to larger.
- Oakbridge-CX · Japan, decommissioned · 4.29 PFlop/s
- Vulcan · USA, decommissioned · 4.29 PFlop/s
- Manzano · USA, operational · 4.28 PFlop/s
- CRONOS · France, operational · 4.3 PFlop/s
- Sugon Advanced Computing System PreE · China, operational · 4.33 PFlop/s
- Pegasus (Tsukuba) · Japan, operational · 4.34 PFlop/s
Related systems
Same site
- Viper-GPU · Germany, operational, 2025
- Raven-GPU · Germany, operational, 2021
- COBRA · Germany, decommissioned, 2018
- Draco · Germany, operational, 2016
- Hydra (MPCDF) · Germany, operational, 2013
- VIP (Garching) · Germany, operational, 2008
Same operator
- Genius (Garching) · Germany, operational, 2007
- MPI/IPP Garching IBM p690 · Germany, operational, 2001
- MPI/IPP Garching Cray T3E · Germany, operational, 1996
- DAIS · Germany, operational
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.