Supercomputers/Systems/Mesabi
USA · operational · hpc
Mesabi
Also known as MSI Mesabi
Operated by University of Minnesota .
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
- 474.07 TFlop/s
- Rpeak
- 614.40 TFlop/s
- Rmax ÷ Rpeak
- 77.2%
- Cores
- 15,360
- Power
- 339.2 kW
- Per watt
- 1.4 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 Mesabi, and those in USA. 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 | 474.07 TFlop/s | #159 of 230 · bottom 32% | #207 of 332 · bottom 39% |
| Energy efficiency | 1.4 GF/W | #126 of 180 · bottom 31% | #95 of 162 · bottom 43% |
| Rmax ÷ Rpeak | 77.2 % | #103 of 226 · top 45% | #140 of 331 · top 42% |
| Power draw | 339 kW | #119 of 182 · bottom 36% | #114 of 171 · bottom 35% |
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 | - HP Apollo 6000 Xl230/250 (MSI: HP Linux distributed cluster) | Hewlett Packard Enterprise | - | Reported | top500.orgweb.archive.org |
| Accelerator | NVIDIA Tesla K20X Accelerators CUDA (Kepler GK110) · 2,688C · 0.732 GHz · 28 nm · 235 W NVIDIA Tesla K20X, two in each of 40 GPU nodes | NVIDIA | 80 reported | Reported | web.archive.org |
| CPU | - Intel Xeon E5-2680v3 12C 2.5GHz Haswell, 741 nodes | Intel | - | Reported | web.archive.orgtop500.org |
| Interconnect | Mellanox InfiniBand FDR Interconnect Fat tree · 56 Gb/s per 4x port · About 1 microsecond Infiniband FDR | Mellanox NVDANASDAQ
rolls up to NVIDIA | - | Reported | top500.org |
The read
Mesabi was the general-purpose Haswell cluster of the Minnesota Supercomputing Institute. MSI's archived page describes an HP Linux cluster of 741 nodes with 17,784 Intel Haswell E5-2680v3 cores and 711 TFlop/s peak, 40 nodes with two Tesla K20X GPUs each (105 TFlop/s more), 16 nodes with 1 TB of RAM, 24 with 256 GB, 616 with 64 GB (67 TB in total) and 32 SSD nodes (15 TB); the name was picked by MSI staff from more than 140 suggestions. TOP500 lists the HP Apollo 6000 Xl230/250 at 15,360 cores, 474.07 TFlop/s Rmax, 614.4 TFlop/s Rpeak and 339.2 kW on five lists, 139th in June 2015 to 449th in June 2017; the two core counts do not reconcile and neither source says why. An earlier MSI machine is recorded as calhoun-msi.
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 Mesabi added from its source pages
Nearby in scale
The machines closest to Mesabi by Rmax, smaller to larger.
- iREMB · South Korea, operational · 0.47 PFlop/s
- Weta Digital Apollo 6000 · New Zealand, upgraded · 0.47 PFlop/s
- Shadow (Mississippi State) · USA, operational · 0.46 PFlop/s
- Shepard · USA, operational · 0.48 PFlop/s
- BlueGene/L · USA, decommissioned · 0.48 PFlop/s
- Vestas1 · Denmark, operational · 0.48 PFlop/s
Related systems
Same operator
- Calhoun (MSI) · USA, operational, 2007
- Agate · USA, operational
Same accelerator
- ILGTI Cray XC30 · India, operational, 2015
- HPC2 · Italy, operational, 2014
- QB-2 · USA, operational, 2014
- Blue Waters · USA, decommissioned, 2013
- Fionn · Ireland, decommissioned, 2013
- Hydra (MPCDF) · Germany, operational, 2013
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.