Parts/Interconnect/Mellanox InfiniBand EDR
Interconnect · EDR · 2014
Mellanox InfiniBand EDR
100 Gb/s InfiniBand, dual-rail in Summit and Sierra. Mellanox was independent at build time; NVIDIA acquired it in 2020.
Why this part
Summit and Sierra both use EDR in a non-blocking fat tree, and both were procured before NVIDIA acquired Mellanox in 2020 for 6.9 billion dollars. That acquisition is the single most consequential corporate event in this dataset: it put the dominant accelerator and the dominant fabric under one owner, and it is why the supplier-rollup logic on this site turns a Mellanox edge from 2017 into an NVIDIA position today.
Systems using this part
- Systems
- 83
- Rmax underneath
- 543.6 PFlop/s
- Units recorded
- 26
1 of 83 edges disclose a quantity
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| Bem 2 2021-06 | Interconnect | Poland | - | - | NVIDIA | - |
| Altair 2021-05 | Interconnect | Poland | - | 3.53 PFlop/s | NVIDIA | #85 → |
| Ares 2021 | Interconnect | Poland | - | 2.34 PFlop/s | Mellanox now NVIDIA | - |
| Dragão 2021 | Interconnect | Brazil | - | 8.98 PFlop/s | NVIDIA | #46 → |
| Tryton Plus 2021 | Interconnect | Poland | - | 1.73 PFlop/s | NVIDIA | #360 → |
| ABCI 2.0 2021 | Interconnect | Japan | - | 22.21 PFlop/s | NVIDIA | #12 → |
| Kelvin2 2020-10 | Interconnect | United Kingdom | - | - | NVIDIA | - |
| Marconi-100 2020-05 | Interconnect | Italy | - | 21.64 PFlop/s | NVIDIA | #9 → |
| Atlas (Petrobras) 2020 | Interconnect | Brazil | - | 4.38 PFlop/s | Mellanox now NVIDIA | #56 → |
| Christofari 2019-12 | Interconnect | Russia | - | 6.67 PFlop/s | NVIDIA | #29 → |
| NA-1 2019-11 | Interconnect | Japan | - | 1.30 PFlop/s | Mellanox now NVIDIA | #2 → |
| NVIDIA DGX SuperPOD (V100) 2019-06 | Interconnect | USA | - | 9.44 PFlop/s | NVIDIA | #22 → |
| Béluga 2019-04 | Interconnect | Canada | - | 2.28 PFlop/s | NVIDIA | #106 → |
| NREL Eagle 2019-01 | Interconnect | USA | - | 4.85 PFlop/s | NVIDIA | #43 → |
| Pangea III 2019 | Interconnect | France | - | 17.86 PFlop/s | NVIDIA | #61 → |
| AiMOS 2019 | Interconnect | USA | - | 8.34 PFlop/s | NVIDIA | #24 → |
| Fênix 2019 | Interconnect | Brazil | - | 3.16 PFlop/s | Mellanox now NVIDIA | #82 → |
| Ogbon 2019 | Interconnect | Brazil | - | 1.60 PFlop/s | NVIDIA | #347 → |
| Longhorn 2019 | Interconnect | USA | - | 2.29 PFlop/s | NVIDIA | #120 → |
| Satori (MIT) 2019 | Interconnect | USA | - | 1.46 PFlop/s | NVIDIA | #373 → |
| Artemis (G42) 2019 | Interconnect | United Arab Emirates | - | 7.26 PFlop/s | NVIDIA | #19 → |
| Computerome 2.0 2019 | Interconnect | Denmark | - | - | NVIDIA | - |
| cHARISMa 2019 | Interconnect | Russia | - | - | NVIDIA | - |
| Hypercluster (Microsoft Research) 2019 | Interconnect | USA | - | 2.67 PFlop/s | NVIDIA | #95 → |
| COBALT 2.0 2019 | Interconnect | Netherlands | 26 | - | NVIDIA | - |
| Circe 2018-11 | Interconnect | USA | - | 3.06 PFlop/s | NVIDIA | #61 → |
| ABCI 2018-08 | Interconnect | Japan | - | 19.88 PFlop/s | NVIDIA | #7 → |
| AFI-NITY 2018-08 | Interconnect | Japan | - | 1.69 PFlop/s | NVIDIA | #126 → |
| Sierra 2018-06 | Interconnect | USA | - | 94.64 PFlop/s | Mellanox now NVIDIA | #62 → |
| Summit 2018-06 | Interconnect | USA | - | 148.6 PFlop/s | Mellanox now NVIDIA | #58 → |
| MareNostrum P9 CTE 2018-05 | Interconnect | Spain | - | 1.02 PFlop/s | Mellanox now NVIDIA | #9 → |
| Niagara 2018-03 | Interconnect | Canada | - | 3.60 PFlop/s | NVIDIA | #70 → |
| HPC4 2018-01 | Interconnect | Italy | - | 12.21 PFlop/s | NVIDIA | #13 → |
| Lassen 2018 | Interconnect | USA | - | 18.20 PFlop/s | NVIDIA | #10 → |
| JUWELS Cluster 2018 | Interconnect | Germany | - | 6.18 PFlop/s | NVIDIA | #171 → |
| Electra 2018 | Interconnect | USA | - | 5.44 PFlop/s | NVIDIA | #46 → |
| ITO Subsystem A 2018 | Interconnect | Japan | - | 4.54 PFlop/s | NVIDIA | #45 → |
| Zhores 2018 | Interconnect | Russia | - | - | NVIDIA | - |
| CMA PAI-B 2018 | Interconnect | China | - | 2.55 PFlop/s | NVIDIA | #68 → |
| Astra 2018 | Interconnect | USA | - | 1.83 PFlop/s | NVIDIA | #204 → |
| ISM SGI 8600 2018 | Interconnect | Japan | - | 941.28 TFlop/s | NVIDIA | #462 → |
| Genius (KU Leuven) 2018 | Interconnect | Belgium | - | - | NVIDIA | - |
| Ansel (LLNL) 2018 | Interconnect | USA | - | 1.29 PFlop/s | Mellanox now NVIDIA | #471 → |
| Quinde I 2017-12-14 | Interconnect | Ecuador | - | 232.00 TFlop/s | Mellanox now NVIDIA | - |
| OCTOPUS 2017-12 | Interconnect | Japan | - | - | NVIDIA | - |
| Brown (Purdue) 2017-10 | Interconnect | USA | - | 790.00 TFlop/s | Mellanox now NVIDIA | #303 → |
| Gyoukou 2017-06 | Interconnect | Japan | - | 19.14 PFlop/s | Mellanox now NVIDIA | #4 → |
| Graham 2017-05 | Interconnect | Canada | - | 1.23 PFlop/s | Mellanox now NVIDIA | #166 → |
| Wilkes-2 2017 | Interconnect | United Kingdom | - | 1.19 PFlop/s | Mellanox now NVIDIA | #101 → |
| AWE Damson 2017 | Interconnect | United Kingdom | - | 3.24 PFlop/s | NVIDIA | #93 → |
| HOKUSAI BigWaterfall 2017 | Interconnect | Japan | - | 1.60 PFlop/s | NVIDIA | #83 → |
| RAIDEN 2017 | Interconnect | Japan | - | 1.21 PFlop/s | NVIDIA | #171 → |
| Reedbush-L 2017 | Interconnect | Japan | - | 810.00 TFlop/s | NVIDIA | #292 → |
| CMA PAI-A 2017 | Interconnect | China | - | 2.44 PFlop/s | NVIDIA | #74 → |
| Centennial (ARL) 2017 | Interconnect | USA | - | 2.44 PFlop/s | NVIDIA | #53 → |
| Joliot-Curie SKL 2017 | Interconnect | France | - | 4.07 PFlop/s | Mellanox now NVIDIA | #34 → |
| Iridis 5 2017 | Interconnect | United Kingdom | - | 860.00 TFlop/s | NVIDIA | #252 → |
| AIST AI Cloud 2017 | Interconnect | Japan | - | 961.00 TFlop/s | Mellanox now NVIDIA | #3 → |
| Suiren2 2017 | Interconnect | Japan | - | 798.00 TFlop/s | Mellanox now NVIDIA | #2 → |
| Facebook DGX-1 Cluster 2017 | Interconnect | USA | - | 3.31 PFlop/s | NVIDIA | #32 → |
| Hellfire 2017 | Interconnect | USA | - | 1.05 PFlop/s | Mellanox now NVIDIA | #122 → |
| Scafell Pike 2017 | Interconnect | United Kingdom | - | 1.83 PFlop/s | Mellanox now NVIDIA | #396 → |
| DGX SaturnV 2016-11 | Interconnect | USA | - | 4.00 PFlop/s | Mellanox now NVIDIA | #29 → |
| GOSAT-2 RCF2 2016-03 | Interconnect | Japan | - | 770.40 TFlop/s | Mellanox now NVIDIA | #8 → |
| Cheyenne 2016 | Interconnect | USA | - | 4.79 PFlop/s | NVIDIA | #198 → |
| Fram 2016 | Interconnect | Norway | - | 950.00 TFlop/s | NVIDIA | #154 → |
| ASPIRE 1 2016 | Interconnect | Singapore | - | 1.01 PFlop/s | Mellanox now NVIDIA | #94 → |
| BrENIAC 2016 | Interconnect | Belgium | - | 548.00 TFlop/s | Mellanox now NVIDIA | #197 → |
| Owens 2016 | Interconnect | USA | - | 705.97 TFlop/s | Mellanox now NVIDIA | #273 → |
| Sakura (PEZY) 2016 | Interconnect | Japan | - | 824.70 TFlop/s | Mellanox now NVIDIA | #3 → |
| Reedbush-U 2016 | Interconnect | Japan | - | 459.83 TFlop/s | NVIDIA | #362 → |
| Cobalt (CCRT) 2016 | Interconnect | France | - | 1.30 PFlop/s | Mellanox now NVIDIA | #470 → |
| Sherlock 2014-01 | Interconnect | USA | - | - | Mellanox now NVIDIA | - |
| HPC Saulėtekis | Interconnect | Lithuania | - | - | NVIDIA | - |
| Taiwania 2 | Interconnect | Taiwan | - | 9.00 PFlop/s | Mellanox now NVIDIA | #20 → |
| Goethe-NHR | Interconnect | Germany | - | 9.09 PFlop/s | NVIDIA | #9 → |
| Bede | Interconnect | United Kingdom | - | - | NVIDIA | - |
| Shoubu System B | Interconnect | Japan | - | 1.06 PFlop/s | NVIDIA | #375 → |
| PARAM Shakti | Interconnect | India | - | - | Mellanox now NVIDIA | - |
| Triton | Interconnect | Finland | - | - | NVIDIA | - |
| CSF3 | Interconnect | United Kingdom | - | - | NVIDIA | - |
| NOAA Mars | Interconnect | USA | - | 1.12 PFlop/s | Mellanox now NVIDIA | #208 → |
| NOAA Venus | Interconnect | USA | - | 1.12 PFlop/s | Mellanox now NVIDIA | #132 → |
“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.