Supercomputers/Systems/Magnolia (Kyoto)
Japan · operational · hpc
Magnolia (Kyoto)
Also known as magnolia Kyoto ACCMS
Operated by Kyoto University ACCMS .
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
- 307.20 TFlop/s
- Rpeak
- 428.65 TFlop/s
- Rmax ÷ Rpeak
- 71.7%
- Cores
- 11,648
- Power
- 165.3 kW
- Per watt
- 1.9 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 Magnolia (Kyoto), 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 | 307.20 TFlop/s | #180 of 245 · bottom 28% | #93 of 131 · bottom 30% |
| Energy efficiency | 1.9 GF/W | #79 of 198 · top 40% | #54 of 74 · bottom 28% |
| Rmax ÷ Rpeak | 71.7 % | #177 of 243 · bottom 28% | #77 of 130 · bottom 42% |
| Power draw | 165 kW | #167 of 199 · bottom 17% | #57 of 74 · bottom 24% |
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 | - Cray XC30/XC40 (TOP500: Cray XC40) | Cray HPENYSE
rolls up to Hewlett Packard Enterprise | - | Reported | top500.orgiimc.kyoto-u.ac.jp |
| CPU | - Intel Xeon E5-2695 v3 14C 2.3GHz, two per node, 416 nodes | Intel | 832 reported | Reported | iimc.kyoto-u.ac.jptop500.org |
| Interconnect | Cray Aries Interconnect Dragonfly · About 10 GB/s per direction · Around 1 microsecond Dragonfly network (TOP500: Aries interconnect) | Cray HPENYSE
rolls up to Hewlett Packard Enterprise | - | Reported | iimc.kyoto-u.ac.jptop500.org |
The read
Magnolia was System D of Kyoto University's Academic Center for Computing and Media Studies, put in service in July 2014. The center's newsletter (Vol. 13 No. 1) describes 416 nodes of two Xeon E5-2695 v3 (Haswell) processors on a Dragonfly network, calling it a Cray XC30, delivering 428 TFlop/s against a 400 TFlop/s contract and 1.03 TFlop/s per node; it had been contracted as a mid-rental expansion of the Camphor, Laurel and Cinnamon procurement and ran beside the Xeon Phi system Camellia. TOP500 lists it as a Cray XC40 with 11,648 cores, 307.2 TFlop/s Rmax, 428.65 TFlop/s Rpeak and 165.3 kW, 174th in November 2014 to 457th in June 2016 (four lists).
Timeline
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Change history
- 2026-10-18 Magnolia (Kyoto) added from its source pages
Nearby in scale
The machines closest to Magnolia (Kyoto) by Rmax, smaller to larger.
- Hercules (ARL) · USA, operational · 0.3 PFlop/s
- Monte Rosa · Switzerland, decommissioned · 0.3 PFlop/s
- Lobachevsky · Russia, operational · 0.29 PFlop/s
- Haeon · South Korea, decommissioned · 0.32 PFlop/s
- Haedam · South Korea, decommissioned · 0.32 PFlop/s
- L-CSC · Germany, operational · 0.32 PFlop/s
Related systems
Same operator
- Camellia (Kyoto) · Japan, operational, 2014
- Camphor 1 · Japan, operational, 2012
- Camphor 2 · Japan, operational, 2016
- Laurel 2 · Japan, operational, 2016
- T2K Kyodai · Japan, operational, 2008
- Camphor 3 · Japan, operational, 2023
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.