Parts/Interconnect/Mellanox InfiniBand QDR
Interconnect · QDR · 2008
Mellanox InfiniBand QDR
40 Gb/s InfiniBand, the generation between DDR and FDR in this taxonomy. Supplied by Mellanox as an independent company before the 2019 NVIDIA acquisition.
Why this part
Fills the one generation gap left in this dataset's InfiniBand ladder between DDR and FDR. It is the fabric TOP500 records for both Nebulae and the Sunway BlueLight MPP, two otherwise unrelated Chinese systems from the same two-year window, which is a reminder that the interconnect layer was one of the first places these machines bought rather than built.
Systems using this part
- Systems
- 47
- Rmax underneath
- 17.88 PFlop/s
- Units recorded
- -
0 of 47 edges disclose a quantity
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| BA-HPC C2 2016-08 | Interconnect | Egypt | - | - | Mellanox now NVIDIA | - |
| Grace (UCL) 2015-12 | Interconnect | United Kingdom | - | 341.30 TFlop/s | NVIDIA | #401 → |
| TUPAC 2015-03 | Interconnect | Argentina | - | - | NVIDIA | - |
| Makman-2 2015 | Interconnect | Saudi Arabia | - | 2.25 PFlop/s | NVIDIA | #28 → |
| Hummel 2015 | Interconnect | Germany | - | 195.62 TFlop/s | NVIDIA | #387 → |
| Faris 2014 | Interconnect | Saudi Arabia | - | 816.58 TFlop/s | NVIDIA | #46 → |
| TEIDE-HPC 2013-11 | Interconnect | Spain | - | 273.97 TFlop/s | Mellanox now NVIDIA | #169 → |
| OCuLUS (Paderborn) 2013-03 | Interconnect | Germany | - | 188.72 TFlop/s | Mellanox now NVIDIA | #174 → |
| Miztli 2013-01 | Interconnect | Mexico | - | 92.28 TFlop/s | Mellanox now NVIDIA | #348 → |
| Makman-1 2013 | Interconnect | Saudi Arabia | - | 441.82 TFlop/s | NVIDIA | #61 → |
| BuTinah 2012-04 | Interconnect | United Arab Emirates | - | 69.36 TFlop/s | NVIDIA | #397 → |
| Xiuhcoatl 2012-01-25 | Interconnect | Mexico | - | - | Mellanox now NVIDIA | - |
| Gordon 2012-01 | Interconnect | USA | - | 218.10 TFlop/s | Mellanox now NVIDIA | #48 → |
| Helios (IFERC) 2012 | Interconnect | Japan | - | 1.24 PFlop/s | Mellanox now NVIDIA | #12 → |
| Curie 2012 | Interconnect | France | - | 1.36 PFlop/s | NVIDIA | #9 → |
| Cab 2012 | Interconnect | USA | - | 347.40 TFlop/s | NVIDIA | #63 → |
| Grifo04 2012 | Interconnect | Brazil | - | 251.50 TFlop/s | NVIDIA | #68 → |
| HA-PACS 2012 | Interconnect | Japan | - | 421.60 TFlop/s | NVIDIA | #41 → |
| Abisko 2012 | Interconnect | Sweden | - | 131.90 TFlop/s | Mellanox now NVIDIA | #130 → |
| Nimbus 2012 | Interconnect | USA | - | 82.69 TFlop/s | NVIDIA | #433 → |
| Formosa 5 2012 | Interconnect | Taiwan | - | 89.94 TFlop/s | NVIDIA | #232 → |
| Faenov (Windows Azure) 2012 | Interconnect | USA | - | 151.30 TFlop/s | NVIDIA | #165 → |
| SuperMike-II 2012 | Interconnect | USA | - | 110.93 TFlop/s | Mellanox now NVIDIA | #250 → |
| HokieSpeed 2011-12 | Interconnect | USA | - | 120.40 TFlop/s | Mellanox now NVIDIA | #147 → |
| Sunway BlueLight MPP 2011-10 | Interconnect | China | - | 795.90 TFlop/s | NVIDIA | #14 → |
| ALPS (NCHC) 2011-07 | Interconnect | Taiwan | - | 177.10 TFlop/s | NVIDIA | #138 → |
| SAGA-220 2011-05 | Interconnect | India | - | - | Mellanox now NVIDIA | - |
| Zin 2011 | Interconnect | USA | - | 773.70 TFlop/s | NVIDIA | #15 → |
| Lonestar 4 2011 | Interconnect | USA | - | 251.80 TFlop/s | NVIDIA | #39 → |
| Guillimin 2011 | Interconnect | Canada | - | 136.30 TFlop/s | NVIDIA | #55 → |
| RWTH Compute Cluster 2011 2011 | Interconnect | Germany | - | 219.84 TFlop/s | Mellanox now NVIDIA | #32 → |
| Dawson2 (UCLA) 2011 | Interconnect | USA | - | 70.28 TFlop/s | Mellanox now NVIDIA | #148 → |
| TSUBAME2.0 2010-11 | Interconnect | Japan | - | 1.19 PFlop/s | NVIDIA | #4 → |
| Tera-100 2010-10 | Interconnect | France | - | 1.05 PFlop/s | NVIDIA | #6 → |
| LOEWE-CSC 2010 | Interconnect | Germany | - | 285.20 TFlop/s | NVIDIA | #22 → |
| Trestles (SDSC) 2010 | Interconnect | USA | - | 67.32 TFlop/s | Mellanox now NVIDIA | #112 → |
| CHEOPS (Cologne) 2010 | Interconnect | Germany | - | 85.90 TFlop/s | Mellanox now NVIDIA | #90 → |
| Vayu (NCI) 2009-11 | Interconnect | Australia | - | 126.40 TFlop/s | Mellanox now NVIDIA | #51 → |
| VSC-1 2009-11 | Interconnect | Austria | - | 35.48 TFlop/s | NVIDIA | #157 → |
| JUROPA 2009 | Interconnect | Germany | - | 274.80 TFlop/s | NVIDIA | #13 → |
| Lomonosov 2009 | Interconnect | Russia | - | 901.90 TFlop/s | NVIDIA | #26 → |
| Tachyon II 2009 | Interconnect | South Korea | - | 274.80 TFlop/s | NVIDIA | #24 → |
| Tsessebe 2009 | Interconnect | South Africa | - | 61.33 TFlop/s | NVIDIA | #313 → |
| Galileu 2009 | Interconnect | Brazil | - | 64.63 TFlop/s | NVIDIA | #76 → |
| Colosse 2009 | Interconnect | Canada | - | 77.17 TFlop/s | NVIDIA | #63 → |
| Nebulae | Interconnect | China | - | 1.27 PFlop/s | NVIDIA | #2 → |
| Mole-8.5 | Interconnect | China | - | 496.50 TFlop/s | NVIDIA | #43 → |
“Units recorded” sums only the edges where a public source states a quantity. It is a floor, not a total, and should never be read as installed base.