Supercomputers/Systems/TSUBAME-KFC
Japan · upgraded · hpc
TSUBAME-KFC
Operated by Institute of Science Tokyo at Institute of Science Tokyo, Ookayama .
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
- 125.20 TFlop/s
- Rpeak
- 217.00 TFlop/s
- Rmax ÷ Rpeak
- 57.7%
- Cores
- 480
- Power
- 27.8 kW
- Per watt
- 4.5 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.
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 | - Supermicro 1U 4x GPU SuperServer | Supermicro | - | Reported | ir.supermicro.comtitech.ac.jp |
| Accelerator | NVIDIA Tesla K20X Accelerators CUDA (Kepler GK110) · 2,688C · 0.732 GHz · 28 nm · 235 W 4 NVIDIA Tesla K20X GPUs per node, 40 nodes | NVIDIA | 160 reported | Verified | titech.ac.jptop500.org |
| CPU | Intel Xeon E5-2620 v2 (Ivy Bridge EP) CPUs x86-64 · 6C · 2.1 GHz 2 Intel Xeon E5-2620 v2 (Ivy Bridge EP) per node, 40 nodes | Intel | 80 reported | Verified | titech.ac.jptop500.org |
| Interconnect | Mellanox InfiniBand FDR Interconnect Fat tree · 56 Gb/s per 4x port · About 1 microsecond FDR InfiniBand | Mellanox NVDANASDAQ
rolls up to NVIDIA | - | Reported | titech.ac.jp |
| Cooling | Single-phase immersion Cooling Full submersion in a dielectric bath · 100 kW per tank and above Oil-submersion cooling (Green Revolution Cooling), plus air-cooling with cooling towers | Institute of Science Tokyo | - | Reported | titech.ac.jpinsidehpc.com |
The read
A 40-node prototype that Tokyo Institute of Technology's GSIC built specifically to test full oil-immersion cooling ahead of TSUBAME3.0, and the first system anywhere to clear 4 GFlops/Watt. It topped the Green500 list in November 2013 and again in June 2014 before GSIC expanded it in place into TSUBAME-KFC/DL, which took second place in November 2015 as a larger machine on the same site. Rmax here is derived from the published efficiency, 4.503 GFlops/Watt, times the published power draw, 27.8 kW, since Tokyo Tech's own release states the ratio and the power but not a standalone Rmax figure; Rpeak is the 217 TFlop/s double-precision peak Tokyo Tech states directly. Tokyo Tech's own announcement and TOP500's independent Green500 list agree on the rank, efficiency and power, so the tier is verified. Operated under the isct company row for continuity with the other TSUBAME rows in this dataset, matching the convention already used for tsubame-3, even though this system predates the 2024 Tokyo Tech and Tokyo Medical and Dental University merger.
Timeline
- 2013 Announced source
- 2013-10 First operational source
- 2013-11 List debut at #1 (November 2013) 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
Institute of Science Tokyo, Ookayama
- Location
- Tokyo, Japan
- Cooling
- Direct liquid cooling
- Commissioned
- 2006
Change history
- 2026-09-24 TSUBAME-KFC added (Tokyo Institute of Technology GSIC, Japan)
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, part and count | Intel Xeon E5-2620 v2 (Ivy Bridge EP), Mellanox InfiniBand FDR · counts: 80 |
| titech.ac.jp | names system and part | Intel Xeon E5-2620 v2 (Ivy Bridge EP), NVIDIA Tesla K20X, Mellanox InfiniBand FDR |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
Power and cooling
Why air ran out at 20 kW a rack, how warm-water cooling removes the chillers, and what PUE conceals.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.