From WikiChip
Difference between revisions of "supercomputer"

(add links)
(→‎Models)
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
{{title|Supercomputer}}
 
{{title|Supercomputer}}
 
A '''supercomputer''' is a [[computer]] with considerably higher compute, memory, storage, and network capabilities than a typical mainstream computer.
 
A '''supercomputer''' is a [[computer]] with considerably higher compute, memory, storage, and network capabilities than a typical mainstream computer.
 +
 +
== Overview ==
 +
A supercomputer is a computer with typically much higher compute capabilities than comprable mainstream computers. It is often built in order to solve challenging problems in science and engineering.
 +
 +
Though historically (1970s-1990s), supercomputers were built from individual, custom CPUs, modern systems comprise a number of individual standardized commercial systems interconnected to form a large complex system through one of the well-known interconnects architectures such as InfiniBand, Ethernet, and Omni-Path.
  
 
== Models ==
 
== Models ==
 +
:;[[supercomputers|Supercomputers]] • [[DoE]]
 +
 
{| border="0" cellpadding="5" width="90%"
 
{| border="0" cellpadding="5" width="90%"
 
|-  
 
|-  
 
|width="30%" valign="top" align="left"|
 
|width="30%" valign="top" align="left"|
−
*[[supercomputers/archer|Archer]]
+
*{{sc|Archer}} • 2014 °
−
*[[supercomputers/archer 2|Archer 2]]
+
*{{sc|Archer 2}} • 2020 °
−
*[[supercomputers/astra|Astra]]
+
*{{sc|Astra}} • 2018
−
*[[supercomputers/aurora|Aurora]] • 2024 (3)
+
*{{sc|Aurora}} • 2023 (3)
−
*[[supercomputers/cielo|Cielo]]
+
*{{sc|Cielo}} • 2010/2011
−
*[[supercomputers/frontera|Frontera]]
+
*{{sc|Frontera}} • 2019
−
*[[supercomputers/hector|Hector]]
+
*{{sc|Hector}} • 2007 °
−
*[[supercomputers/post-k|Post-K]]
+
*{{sc|Post-K}} (K) • 2021
−
*{{sc|Sierra}}
+
*{{sc|Sierra}} • 2018
−
*[[nudt/tianhe|Tianhe]]
+
*[[nudt/tianhe|Tianhe]] ([[NUDT]])
−
*[[zettascaler|ZettaScaler]]
+
*[[zettascaler|ZettaScaler]] • 2017
 
|width="30%" valign="top" align="left"|
 
|width="30%" valign="top" align="left"|
 
:;NNSA/LLNL
 
:;NNSA/LLNL
−
*[[supercomputers/ats-1|ATS-1]] (Trinity)
+
*{{sc|ATS-1}} ({{sc|Trinity}}) • 2017
−
*[[supercomputers/ats-3|ATS-3]] (Crossroads)
+
*{{sc|ATS-3}} ({{sc|Crossroads}}) • 2023
−
*[[supercomputers/el capitan|ATS-4]] (El Capitan) • 2024 (1) <!--Sierra ?-->
+
*{{sc|ATS-4}} ({{sc|El Capitan}}) • 2024 (1) <!--Sierra ?-->
−
*[[supercomputers/ats-5|ATS-5]] (?)
+
*{{sc|ATS-5}} (?) • 2025
  
−
:;[[Oak Ridge National Laboratory|ORNL]] • [[DoE]]
+
:;ORNL
−
* {{sc|OLCF-2}} (Jaguar)
+
* {{sc|OLCF-2}} ({{sc|Jaguar}}) • 2008/2009
−
* {{sc|OLCF-3}} (Titan)
+
* {{sc|OLCF-3}} ({{sc|Titan}}) • 2012
−
*[[supercomputers/summit|OLCF-4]] (Summit)
+
* {{sc|OLCF-4}} ({{sc|Summit}}) • 2018
−
*[[supercomputers/frontier|OLCF-5]] (Frontier)
+
* {{sc|OLCF-5}} ('''{{sc|Frontier}}''') • 2021 (2) ° (TOP2!)
−
*[[supercomputers/olcf-6|OLCF-6]] (Discovery) • 2024  
+
* {{sc|OLCF-6}} ({{sc|Discovery}}) • 2024  
 
|width="30%" valign="top" align="left"|
 
|width="30%" valign="top" align="left"|
 
:;[[National Energy Research Scientific Computing Center|NERSC]]
 
:;[[National Energy Research Scientific Computing Center|NERSC]]
Line 40: Line 47:
 
* {{sc|NERSC-7}} (Edison)
 
* {{sc|NERSC-7}} (Edison)
 
* {{sc|NERSC-8}} (Cori)
 
* {{sc|NERSC-8}} (Cori)
−
* {{sc|NERSC-9}} (Perlmutter)
+
* {{sc|NERSC-9}} ({{sc|Perlmutter}}) • 2020
−
* {{sc|NERSC-10}} (?)
+
* {{sc|NERSC-10}} (?) • 2024
 
|}
 
|}
 +
----
 +
{| border="0" cellpadding="5" width="90%"
 +
|-
 +
|width="30%" valign="top" align="left"|
 +
:;ANL
 +
*'''{{sc|Aurora}}''' • 2023 ° 2025/06 (3) ^ (TOP3!)
 +
*{{sc|Improv}} • 2023
 +
*{{sc|Polaris}} • 2021
 +
*{{sc|Theta}} • 2017
 +
*{{sc|BeBop}} • 2017
 +
*{{sc|Cooley}} • 2015
 +
*{{sc|Mira}} • 2012
 +
*{{sc|Cetus}} • 2012
 +
*{{sc|Vesta}} • 2012
 +
*{{sc|Blues}} • 2012
 +
*{{sc|Magellan}} • 2010
 +
*{{sc|Jazz 2}} • 2009
 +
*{{sc|Intrepid}} • 2007
 +
*{{sc|eServer}} • 2005
 +
*{{sc|Jazz}} • 2002
 +
|width="30%" valign="top" align="left"|
 +
:;NNSA/LLNL
 +
*'''{{sc|El Capitan}}''' ({{sc|ATS-4}}) • 2024 (1!) ° (TOP1!)
 +
*{{sc|Tuolumne}} • 2024
 +
*{{sc|rzAdams}} • 2024
 +
*{{sc|rzVernal}} • 2022
 +
*{{sc|Tioga}} • 2022
 +
*{{sc|Tenaya}} • 2022
 +
*{{sc|Sierra}} • 2018 (20)
 +
*{{sc|Lassen}} • 2018
 +
*{{sc|Ansel}} • 2018
 +
*{{sc|Vulcan}} • 2012
 +
*{{sc|Sequoia}} • 2011
 +
*{{sc|ATS-6}} (?) • 2026 (?)
 +
|width="30%" valign="top" align="left"|
 +
:;NNSA/LANL/SNL
 +
*{{sc|Cielo}} • 2010/2011
 +
*{{sc|Trinity}} ({{sc|ATS-1}}) • 2017
 +
*{{sc|Crossroads}} ({{sc|ATS-3}}) • 2023 (57)
 +
*{{sc|ATS-5}} (?) • 2025 (?)
 +
*{{sc|ATS-7}} (?) • 2027 (?)
  
−
== Overview ==
+
:;NNSA/LANL
−
A supercomputer is a computer with typically much higher compute capabilities than comprable mainstream computers. It is often built in order to solve challenging problems in science and engineering.  
+
*{{sc|Venado}} • 2024
 +
*{{sc|Cerrillos}} • 2009
 +
*{{sc|Roadrunner}} (1-5) • 2008/2009
 +
 
 +
:;SNL
 +
*{{sc|Astra}} • 2018
 +
*{{sc|Mayer}} • 2017
 +
*{{sc|Sullivan}} • 2015
 +
*{{sc|Hammer}} • 2012
 +
|}
 +
----
 +
{| border="0" cellpadding="5" width="60%"
 +
|-
 +
|width="30%" valign="top" align="left"|
 +
:;SNL/NNSA
 +
*{{sc|Mead}} (CTS-2) • 2025
 +
*{{sc|El Dorado}} (Cray) • 2024 • ({{sc|El Capitan}})
 +
*{{sc|Hops}} (CTS-2) • 2024
 +
*{{sc|Stout}} (CTS-2) • 2023
 +
*{{sc|Amber}} (CTS-2) • 2022
 +
*{{sc|Manzano}} (CTS-1) • 2020
 +
*{{sc|Attaway}} (CTS-1) • 2019
 +
*'''{{sc|Astra}}''' (Apollo) • 2018
 +
*{{sc|Eclipse}} (CTS-1) • 2017
 +
|width="30%" valign="top" align="left"|
 +
*{{sc|Cayenne}} (CTS-1) • 2016
 +
*{{sc|Serrano}} (CTS-1) • 2016
 +
*{{sc|Dark Ghost}} (CTS-1) • 2016
 +
*{{sc|Sky Bridge}} (Cray) • 2015
 +
*{{sc|Dark Bridge}} (Xtreme) • 2012
 +
*{{sc|Dark Sand}} (Xtreme) • 2012
 +
*{{sc|Pecos}} (Xtreme) • 2012
 +
*{{sc|Chama}} (Xtreme) • 2011
 +
*{{sc|Red Storm}} (Cray) • 2009
 +
*{{sc|Thunderbird}} (PowerEdge) • 2005
 +
|}
 +
----
 +
=== DoE Laboratories ===
 +
*[[U.S. Department of Energy]] ([[DoE]])
 +
*National Energy Research Scientific Computing Center ([[NERSC]])
 +
*National Nuclear Security Administration (NNSA)
 +
*Lawrence Livermore National Laboratory (LLNL)
 +
*Los Alamos National Laboratory (LANL)
 +
*Oak Ridge National Laboratory (ORNL)
 +
*Sandia National Laboratories (SNL)
 +
*Argonne National Laboratory (ANL)
 +
*Argonne Leadership Computing Facility (ALCF)
  
−
Though historically (1970s-1990s), supercomputers were built from individual, custom CPUs, modern systems comprise a number of individual standardized commercial systems interconnected to form a large complex system through one of the well-known interconnects architectures such as InfiniBand, Ethernet, and Omni-Path.
+
== TOP500 ==
 +
*[https://top500.org/lists/top500/list/2025/06/ TOP500 List] (2025/06)
  
 
<pre>
 
<pre>
−
System Year Vendor Cores Rmax (GFlop/s) Rpeak (GFlop/s)
+
*Rank - System • Cores • Rmax (PFlop/s) • Rpeak (PFlop/s) • Power (kW)
  
−
*Discovery (OLCF-6) - HPE
+
*1 El Capitan - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS, HPE  
−
*Frontier (OLCF-5) - HPE Cray EX235a, AMD Optimized 3rd Generation  
+
:DOE/NNSA/LLNL, United States • 11,039,616 • 1,742.00 • 2,746.38 • 29,581 kW
−
:EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11, HPE Cray OS
+
*2 Frontier - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11,  
−
:2021 HPE 9,066,176 1,353,000,000 2,055,716,990
+
:HPE Cray OS, DOE/SC/Oak Ridge National Laboratory, United States • 9,066,176 • 1,353.00 • 2,055.72 • 24,607
−
*Summit (OLCF-4) - IBM Power System AC922, IBM POWER9 22C 3.07GHz,  
+
*3 Aurora - HPE Cray EX - Intel Exascale Compute Blade, Xeon CPU Max 9470 52C 2.4GHz, Data Centr GPU Max,  
−
:NVIDIA Volta GV100, Dual-rail Mellanox EDR Infiniband
+
:Slingshot-11, DOE/SC/Argonne National Laboratory, United States  • 9,264,128 • 1,012.00 • 1,980.01 • 38,698
−
:2018 IBM 2,414,592 148,600,000 200,794,880
+
*4 JUPITER Booster - BullSequana XH3000, GH Superchip 72C 3GHz, NVIDIA GH200 Superchip, NVIDIA InfiniBand
−
*Frontier TDS - HPE Cray EX235a, AMD Optimized 3rd Generation  
+
:NDR200, RedHat Enterprise Linux, EVIDEN, EuroHPC/FZJ, Germany • 4,801,344 • 793.40 • 930.00 • 13,088
−
:EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11
+
*5 Eagle - Microsoft NDv5, Xeon Platinum 8480C 48C 2GHz, NVIDIA H100, NVIDIA Infiniband NDR,
−
:2021 HPE 120,832 19,200,000 23,105,540
+
Microsoft Azure, United States • 2,073,600 • 561.20 • 846.84
−
*Titan - Cray XK7, Opteron 6274 16C 2.200GHz, Cray Gemini interconnect, NVIDIA K20x
+
*6 HPC6 - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11,
−
:2012 Cray/HPE 560,640 17,590,000 27,112,550
+
:RHEL 8.9, HPE, Eni S.p.A., Italy  • 3,143,520 • 477.90 • 606.97 • 8,461
−
*Jaguar - Cray XK6, Opteron 6274 16C 2.200GHz, Cray Gemini interconnect, NVIDIA 2090
+
*7 Supercomputer Fugaku - Supercomputer Fugaku, A64FX 48C 2.2GHz, Tofu interconnect D, Fujitsu
−
:2009 Cray/HPE 298,592 1,941,000 2,627,609
+
:RIKEN Centr for Computational Science, Japan • 7,630,848 • 442.01 • 537.21 • 29,899
−
*Jaguar - Cray XT5 QC 2.3 GHz
+
*8 Alps - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11, HPE Cray OS,
−
:2008 Cray/HPE 150,152 1,059,000 1,381,400
+
:Swiss National Supercomputing Centr (CSCS), Switzerland • 2,121,600 • 434.90 • 574.84 • 7,124
−
*Jaguar - Cray XT4 QuadCore 2.1 GHz
+
*9 LUMI - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11,
−
:2008 Cray/HPE 30,976 205,000 260,200
+
:HPE , EuroHPC/CSC, Finland • 2,752,704 • 379.70 • 531.51 • 7,107
−
*Jaguar - Cray XT4/XT3
+
*10 Leonardo - BullSequana XH2000, Xeon Platinum 8358 32C 2.6GHz, NVIDIA A100 SXM4 64GB, NVIDIA HDR100
−
:2006 Cray/HPE 23,016 101,700 119,350
+
:Infiniband, EVIDEN, EuroHPC/CINECA, Italy  • 1,824,768 • 241.20 • 306.31 • 7,494
−
*Jaguar - Cray XT3, 2.6 GHz dual Core
+
*11 Isambard-AI phase 2 - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11,
−
:2006 Cray/HPE 10,424 43,480 54,204.8
+
:HPE Cray OS, HPE, University of Bristol, United Kingdom  • 1,028,160 • 216.50 • 278.58
−
*Blue Gene/P Solution
+
*12 Tuolumne - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS,
−
:2007 IBM 8,192 23,862 27,850
+
:DOE/NNSA/LLNL, United States  • 1,161,216 • 208.10 • 288.88 • 3,387
−
*Jaguar - Cray XT3, 2.4 GHz
+
*13 ISEG2 - Nebius NB-D-SR-IESH6, Xeon Platinum 8468 48C 2.1GHz, NVIDIA H200 SXM5 141 GB,
−
:2005 Cray/HPE 5,200 20,527 24,960
+
:Infiniband NDR400, Ubuntu 22.04, Nebius AI, Netherlands  • 718,848 • 202.40 • 338.49 • 5,300
 +
*14 Mare Nostrum 5 ACC - BullSequana XH3000, Xeon Platinum 8460Y+ 32C 2.3GHz, NVIDIA H100 64GB,
 +
:Infiniband NDR, EVIDEN , EuroHPC/BSC, Spain  • 663,040 • 175.30 • 249.44 • 4,159
 +
*15 ABCI 3.0 - HPE Cray XD670, Xeon Platinum 8558 48C 2.1GHz, NVIDIA H200 SXM5 141GB, Infiniband NDR200,
 +
:Institute of Advanced Industrial Science and Technology, Japan • 479,232 • 145.10 • 181.49 • 3,596
 +
*16 Eos NVIDIA DGX SuperPOD - NVIDIA DGX H100, Xeon Platinum 8480C 56C 3.8GHz, NVIDIA H100,  
 +
:Infiniband NDR400, Nvidia, NVIDIA Corporation, United States  • 485,888 • 121.40 • 188.65
 +
*17 Discovery 6 - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11,
 +
:HPE Cray OS, HPE , ExxonMobil, United States  • 822,528 • 118.60 • 222.86
 +
*18 SSC-24 - HPE Cray XD670, Xeon Gold 6430 32C 2.1GHz, NVIDIA H100 SXM5 80GB, Infiniband NDR400,
 +
:RHEL 9.2, HPE , Samsung Electronics, South Korea  • 349,440 • 106.20 • 151.10
 +
*19 Venado - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11,
 +
:HPE, DOE/NNSA/LANL, United States  • 481,440 • 98.51 • 130.44 • 1,662
 +
*20 Sierra - IBM Power System AC922, IBM POWER9 22C 3.1GHz, NVIDIA Volta GV100, Mellanox EDR Infiniband,
 +
:IBM/NVIDIA/Mellanox, DOE/NNSA/LLNL, United States • 1,572,480 • 94.64 • 125.71 • 7,438
 +
*21 Sunway TaihuLight - Sunway MPP, Sunway SW26010 260C 1.45GHz, Sunway, NRCPC
 +
:National Supercomputing Centr in Wuxi, China  • 10,649,600 • 93.01 • 125.44 • 15,371
 +
*22 CHIE-3 - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100, Infiniband NDR400,
 +
:Ubuntu 22.04.4 LTS, Nvidia , SoftBank Corp., Japan • 297,840 • 91.94 • 138.32
 +
*23 CEA-HE - BullSequana XH3000, GH Superchip 72C 3GHz, NVIDIA GH200 Superchip, Quad-Rail BXI v2, Linux,
 +
:EVIDEN , Commissariat a l'Energie Atomique (CEA), France • 548,352 • 90.79 • 171.26 • 1,770
 +
*24 CHIE-2 - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100, Infiniband NDR400,
 +
:Ubuntu 22.04.4 LTS, Nvidia , SoftBank Corp., Japan • 297,840 • 89.78 • 138.32
 +
*25 Perlmutter - HPE Cray EX 235n, AMD EPYC 7763 64C 2.45GHz, NVIDIA A100 SXM4 40 GB, Slingshot-11,
 +
:HPE, DOE/SC/LBNL/NERSC, United States  • 888,832  • 79.23 • 113.00 • 2,945
 +
*26 Njoerd - HPE Cray XD670, Xeon Platinum 8462Y+ 32C 2.8GHz, NVIDIA H100 SXM5 80GB, Infiniband NDR400,
 +
:Ubuntu 22.04, HPE, Northern Data Group, United Kingdom   • 273,280 • 78.20 • 106.28
 +
*27 ABCI-Q - Supermicro SYS-221GE-TNHT-LCC, Intel Xeon Platinum 8558 48C 2.1GHz, NVIDIA H100 SXM5 80GB,
 +
:Infiniband NDR, Rocky Linux 9.4, AIST, Fujitsu, Japan • 315,120  • 74.58 • 99.35 • 1,834
 +
*28 El Dorado - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS,
 +
:Sandia National Laboratories, United States  • 383,040 • 68.02 • 95.29 • 1,110
 +
*29 Gefion - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100 SXM5 80GB, HDR100 Infiniband,  
 +
:Red Hat Enterprise Linux, Danish Centr for AI Innovation, Denmark • 223,088 • 66.59 • 100.63 • 1,753
 +
*30 Selene - NVIDIA DGX A100, AMD EPYC 7742 64C 2.25GHz, NVIDIA A100, Mellanox HDR Infiniband, Nvidia
 +
:NVIDIA Corporation, United States • 555,520  • 63.46 • 79.22 • 2,646
 
</pre>
 
</pre>
 +
 +
== See also ==
 +
*[[AMD]] • [[EPYC]]
 +
*[[Intel]] • [[Xeon]]
 +
*[[IBM]] • [[NVIDIA]]
 +
*[[DoE]] • [[NERSC]]
  
  
 
[[Category:supercomputers]]
 
[[Category:supercomputers]]
 
[[Category:all supercomputers]]
 
[[Category:all supercomputers]]

Latest revision as of 19:04, 30 October 2025

A supercomputer is a computer with considerably higher compute, memory, storage, and network capabilities than a typical mainstream computer.

Overview[edit]

A supercomputer is a computer with typically much higher compute capabilities than comprable mainstream computers. It is often built in order to solve challenging problems in science and engineering.

Though historically (1970s-1990s), supercomputers were built from individual, custom CPUs, modern systems comprise a number of individual standardized commercial systems interconnected to form a large complex system through one of the well-known interconnects architectures such as InfiniBand, Ethernet, and Omni-Path.

Models[edit]

Supercomputers • DoE
NNSA/LLNL
ORNL
NERSC
ANL
NNSA/LLNL
NNSA/LANL/SNL
NNSA/LANL
SNL
SNL/NNSA

DoE Laboratories[edit]

  • U.S. Department of Energy (DoE)
  • National Energy Research Scientific Computing Center (NERSC)
  • National Nuclear Security Administration (NNSA)
  • Lawrence Livermore National Laboratory (LLNL)
  • Los Alamos National Laboratory (LANL)
  • Oak Ridge National Laboratory (ORNL)
  • Sandia National Laboratories (SNL)
  • Argonne National Laboratory (ANL)
  • Argonne Leadership Computing Facility (ALCF)

TOP500[edit]

*Rank -	System	• Cores	• Rmax (PFlop/s) • Rpeak (PFlop/s) • Power (kW)

*1 El Capitan - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS, HPE 
:DOE/NNSA/LLNL, United States	• 11,039,616	• 1,742.00	• 2,746.38	• 29,581 kW
*2 Frontier - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11, 
:HPE Cray OS, DOE/SC/Oak Ridge National Laboratory, United States • 9,066,176 • 1,353.00 • 2,055.72 • 24,607
*3 Aurora - HPE Cray EX - Intel Exascale Compute Blade, Xeon CPU Max 9470 52C 2.4GHz, Data Centr GPU Max, 
:Slingshot-11, DOE/SC/Argonne National Laboratory, United States  • 9,264,128 • 1,012.00 • 1,980.01 • 38,698
*4 JUPITER Booster - BullSequana XH3000, GH Superchip 72C 3GHz, NVIDIA GH200 Superchip, NVIDIA InfiniBand 
:NDR200, RedHat Enterprise Linux, EVIDEN, EuroHPC/FZJ, Germany	• 4,801,344 • 793.40 • 930.00 • 13,088
*5 Eagle - Microsoft NDv5, Xeon Platinum 8480C 48C 2GHz, NVIDIA H100, NVIDIA Infiniband NDR,  
Microsoft Azure, United States	• 2,073,600	• 561.20	• 846.84	
*6 HPC6 - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11, 
:RHEL 8.9, HPE, Eni S.p.A., Italy  • 3,143,520	• 477.90 • 606.97 • 8,461
*7 Supercomputer Fugaku - Supercomputer Fugaku, A64FX 48C 2.2GHz, Tofu interconnect D, Fujitsu 
:RIKEN Centr for Computational Science, Japan	• 7,630,848 • 442.01 • 537.21 • 29,899
*8 Alps - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11, HPE Cray OS,
:Swiss National Supercomputing Centr (CSCS), Switzerland • 2,121,600 • 434.90 • 574.84 • 7,124
*9 LUMI - HPE Cray EX235a, AMD Optimized 3rd Generation EPYC 64C 2GHz, AMD Instinct MI250X, Slingshot-11, 
:HPE , EuroHPC/CSC, Finland	• 2,752,704 • 379.70 • 531.51 • 7,107
*10 Leonardo - BullSequana XH2000, Xeon Platinum 8358 32C 2.6GHz, NVIDIA A100 SXM4 64GB, NVIDIA HDR100 
:Infiniband, EVIDEN, EuroHPC/CINECA, Italy  • 1,824,768 • 241.20 • 306.31 • 7,494
*11 Isambard-AI phase 2 - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11, 
:HPE Cray OS, HPE, University of Bristol, United Kingdom  • 1,028,160 • 216.50 • 278.58	
*12 Tuolumne - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS, 
:DOE/NNSA/LLNL, United States  • 1,161,216 • 208.10 • 288.88 • 3,387
*13 ISEG2 - Nebius NB-D-SR-IESH6, Xeon Platinum 8468 48C 2.1GHz, NVIDIA H200 SXM5 141 GB, 
:Infiniband NDR400, Ubuntu 22.04, Nebius AI, Netherlands  • 718,848 • 202.40 • 338.49 • 5,300
*14 Mare Nostrum 5 ACC - BullSequana XH3000, Xeon Platinum 8460Y+ 32C 2.3GHz, NVIDIA H100 64GB, 
:Infiniband NDR, EVIDEN , EuroHPC/BSC, Spain  • 663,040 • 175.30 • 249.44 • 4,159
*15 ABCI 3.0 - HPE Cray XD670, Xeon Platinum 8558 48C 2.1GHz, NVIDIA H200 SXM5 141GB, Infiniband NDR200,
:Institute of Advanced Industrial Science and Technology, Japan • 479,232 • 145.10 • 181.49 • 3,596
*16 Eos NVIDIA DGX SuperPOD - NVIDIA DGX H100, Xeon Platinum 8480C 56C 3.8GHz, NVIDIA H100, 
:Infiniband NDR400, Nvidia, NVIDIA Corporation, United States  • 485,888 • 121.40 • 188.65	
*17 Discovery 6 - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11, 
:HPE Cray OS, HPE , ExxonMobil, United States   • 822,528 • 118.60 • 222.86	
*18 SSC-24 - HPE Cray XD670, Xeon Gold 6430 32C 2.1GHz, NVIDIA H100 SXM5 80GB, Infiniband NDR400, 
:RHEL 9.2, HPE , Samsung Electronics, South Korea  • 349,440 • 106.20 • 151.10	
*19 Venado - HPE Cray EX254n, NVIDIA Grace 72C 3.1GHz, NVIDIA GH200 Superchip, Slingshot-11, 
:HPE, DOE/NNSA/LANL, United States  • 481,440 • 98.51 • 130.44 • 1,662
*20 Sierra - IBM Power System AC922, IBM POWER9 22C 3.1GHz, NVIDIA Volta GV100, Mellanox EDR Infiniband, 
:IBM/NVIDIA/Mellanox, DOE/NNSA/LLNL, United States • 1,572,480 • 94.64 • 125.71 • 7,438
*21 Sunway TaihuLight - Sunway MPP, Sunway SW26010 260C 1.45GHz, Sunway, NRCPC 
:National Supercomputing Centr in Wuxi, China  • 10,649,600 • 93.01 • 125.44 • 15,371
*22 CHIE-3 - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100, Infiniband NDR400, 
:Ubuntu 22.04.4 LTS, Nvidia , SoftBank Corp., Japan • 297,840 • 91.94 • 138.32	
*23 CEA-HE - BullSequana XH3000, GH Superchip 72C 3GHz, NVIDIA GH200 Superchip, Quad-Rail BXI v2, Linux, 
:EVIDEN , Commissariat a l'Energie Atomique (CEA), France • 548,352 • 90.79 • 171.26 • 1,770
*24 CHIE-2 - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100, Infiniband NDR400, 
:Ubuntu 22.04.4 LTS, Nvidia , SoftBank Corp., Japan • 297,840 • 89.78 • 138.32	
*25 Perlmutter - HPE Cray EX 235n, AMD EPYC 7763 64C 2.45GHz, NVIDIA A100 SXM4 40 GB, Slingshot-11, 
:HPE, DOE/SC/LBNL/NERSC, United States  • 888,832  • 79.23 • 113.00 • 2,945
*26 Njoerd - HPE Cray XD670, Xeon Platinum 8462Y+ 32C 2.8GHz, NVIDIA H100 SXM5 80GB, Infiniband NDR400, 
:Ubuntu 22.04, HPE, Northern Data Group, United Kingdom	  • 273,280 • 78.20 • 106.28	
*27 ABCI-Q - Supermicro SYS-221GE-TNHT-LCC, Intel Xeon Platinum 8558 48C 2.1GHz, NVIDIA H100 SXM5 80GB, 
:Infiniband NDR, Rocky Linux 9.4, AIST, Fujitsu, Japan • 315,120  • 74.58 • 99.35 • 1,834
*28 El Dorado - HPE Cray EX255a, AMD 4th Gen EPYC 24C 1.8GHz, AMD Instinct MI300A, Slingshot-11, TOSS,
:Sandia National Laboratories, United States  • 383,040 • 68.02 • 95.29 • 1,110
*29 Gefion - NVIDIA DGX H100, Xeon Platinum 8480C 56C 2GHz, NVIDIA H100 SXM5 80GB, HDR100 Infiniband, 
:Red Hat Enterprise Linux, Danish Centr for AI Innovation, Denmark • 223,088 • 66.59 • 100.63 • 1,753
*30 Selene - NVIDIA DGX A100, AMD EPYC 7742 64C 2.25GHz, NVIDIA A100, Mellanox HDR Infiniband, Nvidia 
:NVIDIA Corporation, United States • 555,520  • 63.46 • 79.22 • 2,646

See also[edit]

Retrieved from "https://en.wikichip.org/w/index.php?title=supercomputer&oldid=103893"