Supercomputers/Systems/SAKURAONE CYB
Japan · operational · mixed
SAKURAONE CYB
Operated by SAKURA internet at SAKURA internet (Ishikari) .
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
- 13.05 PFlop/s
- Rpeak
- 14.81 PFlop/s
- Rmax ÷ Rpeak
- 88.1%
- Cores
- 63,744
- 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 SAKURAONE CYB, 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 | 13.05 PFlop/s | #101 of 202 · top 50% | #24 of 131 · top 18% |
| Rmax ÷ Rpeak | 88.1 % | #30 of 199 · top 15% | #29 of 130 · top 22% |
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 | - Supermicro SuperServer B200 | Supermicro | - | Reported | top500.orgsakura.ad.jp |
| Accelerator | NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W 48 B200 nodes with eight NVIDIA B200 SXM 180GB GPUs each (service page; TOP500 entry not named) | NVIDIA | 384 estimated | Reported | sakura.ad.jptop500.org |
| CPU | Intel Xeon 6 6900P (Granite Rapids) CPUs x86-64 · 128C · 2 GHz · Intel 3 · 500 W Two Intel Xeon 6960P (72 cores) per node across 48 nodes | Intel | 96 estimated | Reported | sakura.ad.jptop500.org |
| Operating system | Rocky Linux Operating systems RHEL-compatible Linux, community-maintained · Downstream rebuild of Red Hat Enterprise Linux source Rocky Linux 9.6 | Rocky Enterprise Software Foundation | - | Reported | top500.org |
The read
The B200-based entry in SAKURA internet's SAKURAONE family (see sakuraone for the original H100 machine and sakuraone-cyc for the H200 one). TOP500 lists SAKURAONE CYB as a Supermicro SuperServer B200 with Intel Xeon 6960P 72C 2.7 GHz and NVIDIA B200 SXM 180 GB under 'SAKURA Internet Inc. / Prunus Solutions Inc.', installed 2025, with 63,744 cores, Rmax 13.05 and Rpeak 14.81 PFlop/s, HPCG 380.332 TFlop/s and Rocky Linux 9.6; it ranked 136th in November 2025 and 156th in June 2026. SAKURA's managed-HPC service page offers a B200 partition of 48 nodes with eight B200 GPUs (384 GPUs) and two Xeon 6960P per node (6,912 cores); the 384 GPUs at roughly 38 TFlop/s FP64 each would give a peak close to TOP500's 14.81 PFlop/s, but the page does not name the TOP500 entry, so tying the 48-node partition to CYB is our inference. The larger June 2026 entry CYD (37.79 PFlop/s) does not fit that node count and is not recorded here.
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
Change history
- 2026-10-17 Added SAKURAONE CYB
Nearby in scale
The machines closest to SAKURAONE CYB by Rmax, smaller to larger.
- Conrad (SSC) · Canada, operational · 12.91 PFlop/s
- Moira (SSC) · Canada, operational · 12.91 PFlop/s
- Redtail · USA, operational · 12.85 PFlop/s
- Nano 5 · Taiwan, operational · 13.06 PFlop/s
- Olivia · Norway, operational · 13.2 PFlop/s
- JMA PRIMEHPC FX1000 · Japan, operational · 13.37 PFlop/s
Related systems
Same site
- Koukaryoku PHY H200 Container DC · Japan, operational, 2025
- SAKURAONE CYC · Japan, operational, 2025
- SAKURAONE CYD · Japan, operational, 2026
- SAKURA Apollo 6500 V100 Cluster · Japan, operational, 2019
- SAKURAONE · Japan, operational
Same accelerator
- AICR · USA, operational
- Betty · USA, operational, 2025
- CHIE-4 · Japan, operational, 2025
- Data Section Osaka B200 Supercluster · Japan, planned
- Equinox · USA, planned
- Haein Cluster · South Korea, operational, 2025
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 |
|---|---|---|
| sakura.ad.jp | names system, part and count | NVIDIA B200 SXM, Intel Xeon 6 6900P (Granite Rapids) · counts: 384, 96 |
| top500.org | names system, part and count | NVIDIA B200 SXM, Intel Xeon 6 6900P (Granite Rapids), Rocky Linux · counts: 96 |
Checked 2026-10-18 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
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
Programming a supercomputer
The software stack beneath the application, and why the OS is now a sovereignty signal as much as a technical choice.