Supercomputers/Systems/K computer
Japan · decommissioned · hpc
K computer
Deep dive: Fugaku: the CPU-only No. 1 and its successor
Also known as Kei
Operated by RIKEN at RIKEN Center for Computational Science .
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 1540 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
- 10.51 PFlop/s
- Rpeak
- 11.28 PFlop/s
- Rmax ÷ Rpeak
- 93.2%
- Cores
- 705,024
- Power
- 12.66 MW
- Per watt
- 0.8 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 | - Fujitsu | Fujitsu | - | Reported | global.fujitsur-ccs.riken.jp |
| CPU | Fujitsu SPARC64 VIIIfx CPUs SPARC64 · 8C · 2 GHz · 45 nm · 58 W SPARC64 VIIIfx 2.0GHz | Fujitsu | 88,128 reported | Reported | global.fujitsur-ccs.riken.jp |
| Interconnect | Fujitsu Tofu Interconnect Interconnect 6D mesh/torus · 5 GB/s per link per direction Tofu interconnect | Fujitsu | - | Reported | global.fujitsur-ccs.riken.jp |
| Storage | Fujitsu FEFS Storage Lustre-based parallel filesystem · Roughly 150 PB across multiple volumes on Fugaku Fujitsu Exabyte File System (FEFS) based on Lustre | Fujitsu | - | Verified | wiki.lustre.orgvi4io.org |
| Cooling | Direct liquid cooling, warm water Cooling Direct-to-chip cold plates, facility water loop · Removes 90+ percent of node heat to liquid Fujitsu direct water cooling | Fujitsu | - | Reported | global.fujitsur-ccs.riken.jp |
| Cooling | Direct water-cooled LSI (pre-cold-plate era) Cooling Direct water cooling at the chip package Water cooling system for CPU and other LSIs | Fujitsu | - | Reported | aics.riken.jp |
The read
The first system past ten petaflops, and Fujitsu's proof that a CPU-only design with a custom interconnect could still take the top position in the accelerator era. Its efficiency ratio (Rmax over Rpeak above 93%) has rarely been matched by heterogeneous systems since. Decommissioned in 2019 to clear the floor for Fugaku, which reused the site and the architectural argument.
Timeline
- 2006 Announced source
- 2011-06 First operational source
- 2011-11 List debut at #1 (November 2011) source
- 2019-08 Shut down to clear its site at RIKEN for the Fugaku installation source
- 2019-08 Decommissioned 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
RIKEN Center for Computational Science
- Location
- Kobe, Japan
- Cooling
- Direct liquid cooling (warm water)
- Commissioned
- 2010
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 |
|---|---|---|
| global.fujitsu | could not be fetched | - |
| r-ccs.riken.jp | names system and part | Direct liquid cooling, warm water, Direct water-cooled LSI (pre-cold-plate era) |
Checked 2026-10-18 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Power and cooling
Why air ran out at 20 kW a rack, how warm-water cooling removes the chillers, and what PUE conceals.
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.