Supercomputers/Systems/Reedbush-U
Japan · decommissioned · hpc
Reedbush-U
Also known as Reedbush (CPU only)
Operated by University of Tokyo .
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
- 459.83 TFlop/s
- Rpeak
- 508.03 TFlop/s
- Rmax ÷ Rpeak
- 90.5%
- Cores
- 15,120
- Power
- 199.05 kW
- Per watt
- 2.3 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 | - SGI Rackable C2112-4GP3, 420 nodes | Hewlett Packard Enterprise | 420 reported | Verified | top500.orgcc.u-tokyo.ac.jp |
| CPU | Intel Xeon E5-2695 v4 CPUs x86-64 · 18C · 2.1 GHz · 14 nm · 120 W Intel Xeon E5-2695 v4 18C 2.1GHz, two per node | Intel | 840 estimated | Verified | top500.orgcc.u-tokyo.ac.jp |
| Interconnect | Mellanox InfiniBand EDR Interconnect Fat tree · 100 Gb/s per 4x port · About 0.9 microseconds InfiniBand EDR 4x 100 Gbps full-bisection fat tree | NVIDIA | - | Verified | top500.orgcc.u-tokyo.ac.jp |
| Storage | Lustre Storage Parallel filesystem, separate metadata and object servers 5.04 PB Lustre parallel file system shared across Reedbush | Hewlett Packard Enterprise | - | Reported | cc.u-tokyo.ac.jp |
| Operating system | Red Hat Enterprise Linux Operating systems Linux distribution · Open source, commercial support subscription Red Hat Enterprise Linux 7.2 | Red Hat | - | Reported | top500.org |
The read
CPU-only subsystem of the University of Tokyo Information Technology Center's Reedbush complex, alongside the existing Reedbush-H and Reedbush-L GPU subsystems. The center's system page lists 420 SGI Rackable C2112-4GP3 nodes, each with two 18-core Xeon E5-2695v4 and 256GB of memory, a full-bisection InfiniBand EDR fat tree, 508.03 TFlop/s peak and a 5.04 PB Lustre file system. TOP500 lists 15,120 cores, Rmax 459.83 and Rpeak 508.03 TFlop/s at 199.05 kW, ranked 362nd in November 2016 and 469th in June 2017. The center's service page states that Reedbush-U was terminated on 30 June 2020, while Reedbush-H/L ran until 30 November 2021.
Timeline
- 2016 First operational source
- 2016-11 List debut at #362 (November 2016) source
- 2020-06 Decommissioned source
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Change history
- 2026-10-09 Added Reedbush-U
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 | Intel Xeon E5-2695 v4, Mellanox InfiniBand EDR |
| cc.u-tokyo.ac.jp | names system and part | Intel Xeon E5-2695 v4, Mellanox InfiniBand EDR, Lustre |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.
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.