Supercomputers/Systems/Flow Type II
Japan · operational · mixed
Flow Type II
Also known as Flow Type II subsystem; Nagoya University Flow GPU subsystem
Operated by Information Technology Center, Nagoya University at Nagoya University Information Technology 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 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
- 4.88 PFlop/s
- Rpeak
- 7.52 PFlop/s
- Rmax ÷ Rpeak
- 64.9%
- Cores
- 79,560
- 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 Flow Type II, and those in Japan. 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 | 4.88 PFlop/s | #90 of 213 · top 42% | #37 of 131 · top 28% |
| Rmax ÷ Rpeak | 64.9 % | #142 of 209 · bottom 33% | #99 of 130 · bottom 25% |
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 | - Fujitsu PRIMERGY CX2570 M5 | Fujitsu | - | Reported | top500.orgicts.nagoya-u.ac.jp |
| Accelerator | NVIDIA Tesla V100 Accelerators CUDA (Volta GV100) · 5,120C · 1.53 GHz · 12 nm · 300 W NVIDIA Tesla V100 SXM2, 4 per node on 221 nodes | NVIDIA | 884 estimated | Reported | icts.nagoya-u.ac.jptop500.org |
| CPU | - Intel Xeon Gold 6230 20C 2.1GHz, two per node | Intel | - | Reported | top500.orgicts.nagoya-u.ac.jp |
| Interconnect | Mellanox InfiniBand EDR Interconnect Fat tree · 100 Gb/s per 4x port · About 0.9 microseconds Infiniband EDR | NVIDIA | - | Reported | top500.org |
The read
The GPU subsystem of Nagoya University's Flow supercomputer, listed by TOP500 apart from the Fugaku-type Type I subsystem (flow). Nagoya describes 221 Fujitsu PRIMERGY CX2570 M5 nodes with two Xeon CPUs and four NVIDIA V100 each, 33.888 TFlop/s per node and 7.489 PFlop/s in total, 6.4 TB of SSD per node with local shared storage (BeeGFS, BeeOND, NVMesh), and calls it the centre's first large-scale GPU system, aimed at data science, machine learning and AI; computing service for Flow began on 1 July 2020. The V100 count of 884 here is 221 nodes times four, derived from those figures. TOP500 lists Xeon Gold 6230 20-core 2.1 GHz, V100 SXM2 and InfiniBand EDR, Rmax 4.88 and Rpeak 7.52 PFlop/s (Nagoya's 7.489 is the same peak within 0.4 percent), with no power figure; its 79,560 'cores' far exceed the 8,840 CPU cores of 221 dual-20-core nodes, so TOP500's count includes accelerator cores (inferred). It debuted at 58th in November 2020, ranked 180th in June 2024 (the entry recorded here) and 282nd in November 2025, its last listing. A Nagoya slide deck from 2025 gives March 2026 as the end of Flow's operation and a successor, 'Flow NEXT' (name marked TBD), from 1 October 2026; no source confirms the switch-off, so the row stays operational as of its last list.
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
Nagoya University Information Technology Center
- Location
- Nagoya, Japan
Change history
- 2026-11-01 Added Flow Type II (Nagoya University)
Nearby in scale
The machines closest to Flow Type II by Rmax, smaller to larger.
- Joule 3.0 · USA, operational · 4.86 PFlop/s
- NREL Eagle · USA, operational · 4.85 PFlop/s
- Cheyenne · USA, decommissioned · 4.79 PFlop/s
- BioHive-1 · USA, operational · 4.9 PFlop/s
- CUBOX Cray XD6500 · South Korea, operational · 4.96 PFlop/s
- Cardinal · USA, operational · 5.04 PFlop/s
Related systems
Same site
- Flow · Japan, operational, 2020
- Nagoya FX100 · Japan, decommissioned, 2015
- Nagoya FX1 · Japan, operational, 2009
Same accelerator
- Atlas (Petrobras) · Brazil, operational, 2020
- Dammam-7 · Saudi Arabia, operational, 2020
- Expanse · USA, operational, 2020
- Gadi · Australia, operational, 2020
- Garrawarla · Australia, operational, 2020
- HPC5 · Italy, operational, 2020
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.