Supercomputers/Systems/ABCI
Japan · decommissioned · ai training
ABCI
Also known as AI Bridging Cloud Infrastructure
Operated by AIST at ABCI Data Center, Kashiwa .
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: 1263 of 1345 systems have a readable citation that names them, and 56 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
- 19.88 PFlop/s
- Rpeak
- 32.58 PFlop/s
- Rmax ÷ Rpeak
- 61.0%
- Cores
- 391,680
- Power
- 1.65 MW
- Per watt
- 12.1 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 PRIMERGY CX2570 M4 | Fujitsu | - | Reported | abci.ai |
| Accelerator | NVIDIA Tesla V100 Accelerators CUDA (Volta GV100) · 5,120C · 1.53 GHz · 12 nm · 300 W NVIDIA Volta GV100 | NVIDIA | - | Reported | abci.ai |
| CPU | Intel Xeon Gold 6148 CPUs x86-64 · 20C · 2.4 GHz · 14 nm · 150 W Xeon Gold 6148 20C 2.4GHz | Intel | - | Reported | abci.ai |
| Interconnect | Mellanox InfiniBand EDR Interconnect Fat tree · 100 Gb/s per 4x port · About 0.9 microseconds compute nodes(V) are equipped with 2 ports of InfiniBand EDR | NVIDIA | - | Verified | docs.abci.aitop500.org |
| Cooling | Direct liquid cooling, warm water Cooling Direct-to-chip cold plates, facility water loop · Removes 90+ percent of node heat to liquid warm-water direct liquid cooling with hot-water loop | Fujitsu | - | Reported | abci.ai |
| Operating system | CentOS Operating systems Linux distribution, RHEL-compatible rebuild · Open source, community-maintained Change the OS for compute nodes (V) and interactive nodes (V) from CentOS 7 to Rocky Linux 8 | The CentOS Project | - | Reported | docs.abci.ai |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 4,352 accelerators | 2024-10 | Operator stated | ABCI 1.0 partition: 1,088 nodes with 4,352 NVIDIA V100. The ABCI 2.0 A100 partition (120 nodes, 960 A100) is excluded. Both partitions retired 2024-10-31. | docs.abci.ai |
The read
Built in 2018 explicitly as an AI facility rather than an HPC facility, years before that distinction became standard, and one of the few systems designed that way that did submit an HPL result. Its reported PUE near 1.1 with a hot-water loop makes it one of the better-documented facility rows we hold. Fujitsu, the builder, independently states the AIST tender, the PRIMERGY CX2570 M4 nodes with four Tesla V100 each and a 19.88 PFlop/s LINPACK result, which matches; that second source is why the tier is verified. Fujitsu's release says Xeon Scalable rather than Gold 6148, so the exact CPU SKU rests on the operator.
Timeline
- 2017 Announced source
- 2018-08 First operational source
- 2018-11 List debut at #7 (November 2018) source
- 2024-10-31 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
ABCI Data Center, Kashiwa
- Location
- Kashiwa, Japan
- Cooling
- Direct liquid cooling with hot-water loop
- PUE
- 1.1
- Commissioned
- 2018
Change history
- 2026-10-02 Marked ABCI decommissioned
- 2026-09-19 Raised to verified on an independent second source
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 |
|---|---|---|
| abci.ai | names system and part | Direct liquid cooling, warm water |
| info.archives.global.fujitsu | names system and part | NVIDIA Tesla V100, Direct liquid cooling, warm water, Intel Xeon Gold 6148 |
Checked 2026-10-13 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
Power and cooling
Why air ran out at 20 kW a rack, how warm-water cooling removes the chillers, and what PUE conceals.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.