Parts/Interconnect/Mellanox InfiniBand FDR
Interconnect · FDR · 2011
Mellanox InfiniBand FDR
56 Gb/s InfiniBand, the generation between QDR and EDR and the fabric under a large share of the 2012-2016 installed base.
Why this part
FDR moved to 64b/66b encoding, so 56 Gb/s of signalling delivered about 54 of data instead of the 80 percent that the older 8b/10b scheme allowed. Pleiades at NASA Ames runs it across a hypercube rather than the usual fat tree, one of the few large InfiniBand installations that does, because the machine grew by accretion over more than a decade and a hypercube extends without rewiring what already exists.
Systems using this part
- Systems
- 107
- Rmax underneath
- 96.52 PFlop/s
- Units recorded
- -
0 of 107 edges disclose a quantity
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| Reedbush-H 2017 | Interconnect | Japan | - | 800.00 TFlop/s | NVIDIA | #204 → |
| Kukai 2017 | Interconnect | Japan | - | 460.70 TFlop/s | Mellanox now NVIDIA | #2 → |
| CNIC Era-AI 2017 | Interconnect | China | - | 1.24 PFlop/s | NVIDIA | #110 → |
| Lengau 2016-06 | Interconnect | South Africa | - | 1.03 PFlop/s | NVIDIA | #122 → |
| Pangea II 2016-06 | Interconnect | France | - | 5.28 PFlop/s | NVIDIA | #11 → |
| Draco 2016-05 | Interconnect | Germany | - | 790.70 TFlop/s | NVIDIA | #160 → |
| Phoenix (Adelaide) 2016-02 | Interconnect | Australia | - | 317.50 TFlop/s | NVIDIA | #439 → |
| Vestas1 2016 | Interconnect | Denmark | - | 481.79 TFlop/s | Mellanox now NVIDIA | #442 → |
| Biowulf 2016 | Interconnect | USA | - | 1.97 PFlop/s | NVIDIA | #217 → |
| Tera-1000-1 2016 | Interconnect | France | - | 1.87 PFlop/s | Mellanox now NVIDIA | #45 → |
| Volvo Cars BL460c Cluster 2016 | Interconnect | Sweden | - | 861.55 TFlop/s | NVIDIA | #140 → |
| CARL and EDDY 2016 | Interconnect | Germany | - | 457.23 TFlop/s | NVIDIA | #364 → |
| BinAC 2016 | Interconnect | Germany | - | 334.80 TFlop/s | NVIDIA | #414 → |
| Finis Terrae II 2016 | Interconnect | Spain | - | 213.00 TFlop/s | NVIDIA | - |
| JAEA SGI ICE X 2015-11 | Interconnect | Japan | - | 1.93 PFlop/s | NVIDIA | #35 → |
| Salomon 2015-09-15 | Interconnect | Czechia | - | 1.46 PFlop/s | Mellanox now NVIDIA | #40 → |
| DKRZ Mistral 2015-07 | Interconnect | Germany | - | 3.01 PFlop/s | NVIDIA | #73 → |
| Sekirei (ISSP System B) 2015-07 | Interconnect | Japan | - | 1.18 PFlop/s | NVIDIA | #54 → |
| SuperMUC Phase 2 2015-06 | Interconnect | Germany | - | 2.81 PFlop/s | NVIDIA | #21 → |
| ARIS 2015-06 | Interconnect | Greece | - | 170.00 TFlop/s | NVIDIA | #468 → |
| CoolMUC-2 2015-05 | Interconnect | Germany | - | 366.36 TFlop/s | NVIDIA | #262 → |
| Comet 2015-04 | Interconnect | USA | - | - | Mellanox now NVIDIA | - |
| Shirokane3 2015-04 | Interconnect | Japan | - | 342.58 TFlop/s | NVIDIA | #400 → |
| Tryton 2015-03 | Interconnect | Poland | - | 1.01 PFlop/s | Mellanox now NVIDIA | #93 → |
| Abacus 2.0 2015-03 | Interconnect | Denmark | - | 288.90 TFlop/s | NVIDIA | #243 → |
| Prometheus 2015 | Interconnect | Poland | - | 1.67 PFlop/s | NVIDIA | #49 → |
| Shoubu 2015 | Interconnect | Japan | - | 610.00 TFlop/s | NVIDIA | #136 → |
| Taurus 2015 | Interconnect | Germany | - | 1.03 PFlop/s | Mellanox now NVIDIA | #66 → |
| Earth System Numerical Simulator 2015 | Interconnect | China | - | 738.00 TFlop/s | NVIDIA | #182 → |
| XStream 2015 | Interconnect | USA | - | 781.30 TFlop/s | Mellanox now NVIDIA | #123 → |
| P&G Apollo 6000 Cluster 2015 | Interconnect | USA | - | 233.99 TFlop/s | NVIDIA | #440 → |
| Intel SC D2P4 2015 | Interconnect | USA | - | 830.00 TFlop/s | NVIDIA | #81 → |
| PSNC Eagle 2015 | Interconnect | Poland | - | 1.01 PFlop/s | NVIDIA | #80 → |
| ABACUS I 2015 | Interconnect | Mexico | - | 277.50 TFlop/s | NVIDIA | #255 → |
| Thunder (AFRL) 2015 | Interconnect | USA | - | 3.13 PFlop/s | NVIDIA | #19 → |
| Sango 2015 | Interconnect | Japan | - | 252.70 TFlop/s | NVIDIA | #403 → |
| Iceberg 2015 | Interconnect | USA | - | 384.34 TFlop/s | NVIDIA | #254 → |
| Avitohol 2015 | Interconnect | Bulgaria | - | 264.21 TFlop/s | NVIDIA | #389 → |
| BURA 2015 | Interconnect | Croatia | - | 233.56 TFlop/s | NVIDIA | #441 → |
| BEM 2015 | Interconnect | Poland | - | 695.59 TFlop/s | NVIDIA | #128 → |
| Topaz (ERDC) 2015 | Interconnect | USA | - | 3.32 PFlop/s | Mellanox now NVIDIA | - |
| QPACE2 2015 | Interconnect | Germany | - | 206.40 TFlop/s | Mellanox now NVIDIA | #500 → |
| Cooley 2015 | Interconnect | USA | - | 240.40 TFlop/s | Mellanox now NVIDIA | #298 → |
| Piz Kesch 2015 | Interconnect | Switzerland | - | 282.50 TFlop/s | NVIDIA | #353 → |
| L-CSC 2014-11 | Interconnect | Germany | - | 316.70 TFlop/s | Mellanox now NVIDIA | #1 → |
| ForHLR I 2014-09 | Interconnect | Germany | - | 206.30 TFlop/s | Mellanox now NVIDIA | #243 → |
| CNIC Era 2014-06 | Interconnect | China | - | 406.80 TFlop/s | NVIDIA | #240 → |
| Savio 2014-05 | Interconnect | USA | - | - | Mellanox now NVIDIA | - |
| COMA (PACS-IX) 2014-04-15 | Interconnect | Japan | - | 746.00 TFlop/s | NVIDIA | #52 → |
| Lomonosov-2 2014 | Interconnect | Russia | - | 2.48 PFlop/s | NVIDIA | - |
| HPC2 2014 | Interconnect | Italy | - | 3.19 PFlop/s | Mellanox now NVIDIA | #12 → |
| Polytechnic RSC Tornado 2014 | Interconnect | Russia | - | 658.11 TFlop/s | NVIDIA | #81 → |
| Yemoja 2014 | Interconnect | Brazil | - | 405.36 TFlop/s | NVIDIA | #96 → |
| Occigen 2014 | Interconnect | France | - | 1.63 PFlop/s | Mellanox now NVIDIA | #27 → |
| SuperMIC 2014 | Interconnect | USA | - | 557.03 TFlop/s | Mellanox now NVIDIA | #116 → |
| Lobachevsky 2014 | Interconnect | Russia | - | 289.50 TFlop/s | NVIDIA | #189 → |
| QB-2 2014 | Interconnect | USA | - | 1.05 PFlop/s | Mellanox now NVIDIA | #47 → |
| Porthos 2014 | Interconnect | France | - | 506.36 TFlop/s | NVIDIA | #123 → |
| Airbus HPC4 2014 | Interconnect | France | - | 516.90 TFlop/s | NVIDIA | #73 → |
| Prolix 2014 | Interconnect | France | - | 2.17 PFlop/s | Mellanox now NVIDIA | - |
| ISM Supercomputer System I 2014 | Interconnect | Japan | - | 360.00 TFlop/s | NVIDIA | #265 → |
| Constance 2014 | Interconnect | USA | - | 203.24 TFlop/s | NVIDIA | #371 → |
| Lichtenberg I Phase 2 2014 | Interconnect | Germany | - | 249.99 TFlop/s | NVIDIA | #232 → |
| Euler I-II 2014 | Interconnect | Switzerland | - | 237.40 TFlop/s | NVIDIA | #307 → |
| Computerome 1 2014 | Interconnect | Denmark | - | 410.78 TFlop/s | NVIDIA | #121 → |
| QUARTETTO 2013-12-27 | Interconnect | Japan | - | 1.02 PFlop/s | Mellanox now NVIDIA | #37 → |
| TSUBAME-KFC 2013-10 | Interconnect | Japan | - | 125.20 TFlop/s | Mellanox now NVIDIA | #1 → |
| Hydra (MPCDF) 2013-10 | Interconnect | Germany | - | 1.28 PFlop/s | NVIDIA | #28 → |
| Conte 2013-07-30 | Interconnect | USA | - | 976.76 TFlop/s | NVIDIA | #34 → |
| Cartesius 2013-06 | Interconnect | Netherlands | - | 1.09 PFlop/s | NVIDIA | #46 → |
| Spirit (AFRL) 2013-06 | Interconnect | USA | - | 1.42 PFlop/s | NVIDIA | #15 → |
| Lichtenberg I 2013-06 | Interconnect | Germany | - | 213.78 TFlop/s | NVIDIA | #148 → |
| PARAM Yuva II 2013-02 | Interconnect | India | - | 388.44 TFlop/s | NVIDIA | #70 → |
| Stampede 2013-01 | Interconnect | USA | - | 5.17 PFlop/s | Mellanox now NVIDIA | #6 → |
| Aaditya 2013 | Interconnect | India | - | 719.22 TFlop/s | Mellanox now NVIDIA | #37 → |
| Wilkes 2013 | Interconnect | United Kingdom | - | 239.90 TFlop/s | NVIDIA | #167 → |
| Argus 2013 | Interconnect | Canada | - | 136.22 TFlop/s | NVIDIA | #377 → |
| Cascade (PNNL) 2013 | Interconnect | USA | - | 2.54 PFlop/s | NVIDIA | #14 → |
| Athos 2013 | Interconnect | France | - | 352.67 TFlop/s | NVIDIA | #92 → |
| HPC1 2013 | Interconnect | Italy | - | 454.05 TFlop/s | NVIDIA | #70 → |
| Laure 2013 | Interconnect | France | - | 271.75 TFlop/s | NVIDIA | #110 → |
| Grifo06 2013 | Interconnect | Brazil | - | 160.30 TFlop/s | NVIDIA | #220 → |
| Fionn 2013 | Interconnect | Ireland | - | 140.45 TFlop/s | NVIDIA | #359 → |
| Beaufix 2013 | Interconnect | France | - | 2.16 PFlop/s | Mellanox now NVIDIA | - |
| Yellowstone 2012-12 | Interconnect | USA | - | 1.26 PFlop/s | Mellanox now NVIDIA | #13 → |
| Keeneland 2012-11 | Interconnect | USA | - | 319.56 TFlop/s | NVIDIA | #75 → |
| Raijin (NCI) 2012-11 | Interconnect | Australia | - | 978.60 TFlop/s | NVIDIA | #24 → |
| Midway (UChicago) 2012-11 | Interconnect | USA | - | 130.16 TFlop/s | Mellanox now NVIDIA | #303 → |
| CRIEPI SGI ICE X 2012-07 | Interconnect | Japan | - | 582.10 TFlop/s | NVIDIA | #39 → |
| SuperMUC 2012-06 | Interconnect | Germany | - | 2.90 PFlop/s | NVIDIA | #4 → |
| SANAM 2012 | Interconnect | Saudi Arabia | - | 532.60 TFlop/s | NVIDIA | #52 → |
| MVS-10P 2012 | Interconnect | Russia | - | 375.70 TFlop/s | NVIDIA | #59 → |
| Triolith 2012 | Interconnect | Sweden | - | 407.24 TFlop/s | Mellanox now NVIDIA | #95 → |
| Vilje 2012 | Interconnect | Norway | - | 396.70 TFlop/s | Mellanox now NVIDIA | #56 → |
| Abel (UiO) 2012 | Interconnect | Norway | - | 178.60 TFlop/s | Mellanox now NVIDIA | #134 → |
| Virgo (IIT Madras) 2012 | Interconnect | India | - | 91.13 TFlop/s | NVIDIA | #224 → |
| Haise 2012 | Interconnect | USA | - | 327.35 TFlop/s | NVIDIA | #73 → |
| Kilrain 2012 | Interconnect | USA | - | 327.35 TFlop/s | NVIDIA | #72 → |
| Cernan 2012 | Interconnect | USA | - | 78.11 TFlop/s | NVIDIA | #469 → |
| Hercules (ARL) 2012 | Interconnect | USA | - | 303.96 TFlop/s | NVIDIA | #95 → |
| Pershing 2012 | Interconnect | USA | - | 350.73 TFlop/s | NVIDIA | #79 → |
| Carter (Purdue) 2011-11 | Interconnect | USA | - | 186.90 TFlop/s | Mellanox now NVIDIA | #54 → |
| Pleiades 2008-11 | Interconnect | USA | - | 5.95 PFlop/s | Mellanox now NVIDIA | #3 → |
| Della | Interconnect | USA | - | - | Mellanox now NVIDIA | - |
| Maia (NAS) | Interconnect | USA | - | 212.91 TFlop/s | Mellanox now NVIDIA | #117 → |
| CSIRO GPU Cluster | Interconnect | Australia | - | 335.30 TFlop/s | NVIDIA | #146 → |
| Bluecrab | Interconnect | USA | - | - | Mellanox now NVIDIA | - |
“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.