TOP500

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

TOP500 project logo

The TOP500 project ranks and details the 500 most powerful non-distributed computer systems in the world. The project was started in 1993 and publishes an updated list of the supercomputers twice a year. The first of these updates always coincides with the International Supercomputing Conference in June, and the second is presented at the ACM/IEEE Supercomputing Conference in November. The project aims to provide a reliable basis for tracking and detecting trends in high-performance computing and bases rankings on HPL,[1] a portable implementation of the high-performance LINPACK benchmark written in Fortran for distributed-memory computers.

China currently dominates the list with 229 supercomputers, leading the second place (United States) by a record margin of 121. Since June 2018, the American Summit is the world's most powerful supercomputer, reaching 143.5 petaFLOPS on the LINPACK benchmarks.

The TOP500 list is compiled by Jack Dongarra of the University of Tennessee, Knoxville, Erich Strohmaier and Horst Simon of the National Energy Research Scientific Computing Center (NERSC) and Lawrence Berkeley National Laboratory (LBNL), and from 1993 until his death in 2014, Hans Meuer of the University of Mannheim, Germany.

History[edit]

Rapid growth of supercomputer performance, based on data from top500.org site. The logarithmic y-axis shows performance in GFLOPS.
  Combined performance of 500 largest supercomputers
  Fastest supercomputer
  Supercomputer in 500th place

In the early 1990s, a new definition of supercomputer was needed to produce meaningful statistics. After experimenting with metrics based on processor count in 1992, the idea arose at the University of Mannheim to use a detailed listing of installed systems as the basis. In early 1993, Jack Dongarra was persuaded to join the project with his LINPACK benchmarks. A first test version was produced in May 1993, partly based on data available on the Internet, including the following sources:[2][3]

  • "List of the World's Most Powerful Computing Sites" maintained by Gunter Ahrendt[4]
  • David Kahaner, the director of the Asian Technology Information Program (ATIP);[5] published a report in 1992, titled "Kahaner Report on Supercomputer in Japan"[3] which had an immense amount of data.[citation needed]

The information from those sources was used for the first two lists. Since June 1993, the TOP500 is produced bi-annually based on site and vendor submissions only.

Since 1993, performance of the No. 1 ranked position has grown steadily in accordance with Moore's law, doubling roughly every 14 months. As of June 2018, Summit was fastest with an Rpeak[6] of 187.6593 PFLOPS. For comparison, this is over 1,432,513 times faster than the Connection Machine CM-5/1024 (1,024 cores), which was the fastest system in November 1993 (twenty-five years prior) with an Rpeak of 131.0 GFLOPS.[7]

Architecture and operating systems[edit]

In June 2016, a Chinese computer made the top based on SW26010 processors, a new, radically modified, model in the Sunway (or ShenWei) line.

As of June 2018, TOP500 supercomputers are now all 64-bit, mostly based on x86-64 CPUs (Intel EMT64 and AMD AMD64 instruction set architecture), with few exceptions (all based on reduced instruction set computing (RISC) architectures) including 15 supercomputers based on Power Architecture used by IBM POWER microprocessors, six SPARC (all with Fujitsu-designed SPARC-chips, one of which—the K computer—surprisingly made the top in 2011 without a GPU, currently ranked 16th, while only dropping to 3rd on the HPCG list[8]), and two, seemingly related, Chinese designs: the ShenWei-based (ranked 11 in 2011, ranked 158th in November 2016) and Sunway SW26010-based ranked 1 in 2016, making up the remainder (another non-US design is PEZY-SC, while it is an accelerator paired with Intel's Xeon). Before the ascendance of 32-bit x86 and later 64-bit x86-64 in the early 2000s, a variety of RISC processor families made up most TOP500 supercomputers, including RISC architectures such as SPARC, MIPS, PA-RISC, and Alpha.

In recent years heterogeneous computing, mostly using Nvidia's graphics processing units (GPU) as coprocessors, has become a popular way to reach a better performance per watt ratio and higher absolute performance; it is almost required for good performance and to make the top (or top 10), with some exceptions, such as the mentioned SPARC computer without any coprocessors. An x86-based coprocessor, Xeon Phi, has also been used.

Share of processor architecture families in TOP500 supercomputers by time trend

All the fastest supercomputers in the decade since the Earth Simulator supercomputer have used operating systems based on Linux. Since November 2017, all the listed supercomputers (100% of the performance share) use an operating system based on the Linux kernel.[9][10]

Since November 2015, no computer on the list runs Windows. In November 2014, Windows Azure[11] cloud computer was no longer on the list of fastest supercomputers (its best rank was 165 in 2012), leaving the Shanghai Supercomputer Center's Magic Cube as the only Windows-based supercomputer on the list, until it also dropped off the list. It had been ranked 436 in its last appearance on the list released in June 2015, while its best rank was 11 in 2008.[12]

It has been well over a decade since MIPS-based systems (meaning used as host CPUs) dropped entirely off the list[13] but the Gyoukou supercomputer that jumped to 4th place in November 2017 (after a huge upgrade) has MIPS as a small part of the coprocessors. Use of 2,048-core coprocessors (plus 8× 6-core MIPS, for each, that "no longer require to rely on an external Intel Xeon E5 host processor"[14]) make the supercomputer much more energy efficient than the other top 10 (i.e. it is 5th on Green500 and other such ZettaScaler-2.2-based systems take first three spots).[15] At 19.86 million cores, it is by far the biggest system: almost double that of the best manycore system in the TOP500, the Chinese Sunway TaihuLight, currently ranked 3rd.

Top 500 ranking[edit]

Top 10 positions of the 52nd TOP500 in November 2018[16]
Rank Rmax
Rpeak
(PFLOPS)
Name Model Processor Interconnect Vendor Site
country, year
Operating system
1 Steady 143.500
200.795
Summit Power System AC922 POWER9, Tesla V100 Infiniband EDR IBM Oak Ridge National Laboratory
 United States, 2018
Linux (RHEL)
2 Increase 94.640
125.436
Sierra Power System S922LC POWER9, Tesla V100 Infiniband EDR IBM Lawrence Livermore National Laboratory
 United States, 2018
Linux (RHEL)
3 Decrease 93.015
125.436
Sunway TaihuLight Sunway MPP SW26010 Sunway[17] NRCPC National Supercomputing Center in Wuxi
 China, 2016[17]
Linux (Raise)
4 Decrease 61.445
100.679
Tianhe-2A TH-IVB-FEP Xeon E5–2692 v2, Matrix-2000 TH Express-2 NUDT National Supercomputing Center in Guangzhou
 China, 2013
Linux (Kylin)
5 Increase 21.230
27.154
Piz Daint Cray XC50 Xeon E5-2690 v3, Tesla P100 Aries Cray Swiss National Supercomputing Centre
  Switzerland, 2016
Linux (CLE)
6 Decrease 20.159
41.461
Trinity Cray XC40 Xeon E5–2698 v3, Xeon Phi 7250 Aries Cray Los Alamos National Laboratory
 United States, 2015
Linux (CLE)
7 Decrease 19.880
32.577
AI Bridging Cloud Infrastructure[18] PRIMERGY CX2550 M4 Xeon Gold 6148, Tesla V100 Infiniband EDR Fujitsu National Institute of Advanced Industrial Science and Technology
 Japan, 2018
Linux
8 Increase 19.477
26.874
SuperMUC-NG[19] ThinkSystem SD530 Xeon Platinum 8174 Intel Omni-Path Lenovo Leibniz Supercomputing Centre
 Germany, 2018
Linux (SLES)
9 Decrease 17.590
27.113
Titan Cray XK7 Opteron 6274, Tesla K20X Gemini Cray Oak Ridge National Laboratory
 United States, 2012
Linux (CLE, SLES based)
10 Decrease 17.173
20.133
Sequoia Blue Gene/Q A2 Custom IBM Lawrence Livermore National Laboratory
 United States, 2013
Linux (RHEL and CNK)

Legend:

  • Rank – Position within the TOP500 ranking. In the TOP500 list table, the computers are ordered first by their Rmax value. In the case of equal performances (Rmax value) for different computers, the order is by Rpeak. For sites that have the same computer, the order is by memory size and then alphabetically.
  • Rmax – The highest score measured using the LINPACK benchmarks suite. This is the number that is used to rank the computers. Measured in quadrillions of floating point operations per second, i.e., petaFLOPS.
  • Rpeak – This is the theoretical peak performance of the system. Measured in PFLOPS.
  • Name – Some supercomputers are unique, at least on its location, and are thus named by their owner.
  • Model – The computing platform as it is marketed.
  • Processor – The instruction set architecture or processor microarchitecture.
  • Interconnect – The interconnect between computing nodes.
  • Vendor – The manufacturer of the platform and hardware.
  • Site – The name of the facility operating the supercomputer.
  • Country – The country in which the computer is located.
  • Year – The year of installation or last major update.
  • Operating system – The operating system that the computer uses.

Other rankings[edit]

Top countries[edit]

Numbers below represent the number of computers in the TOP500 that are in each of the listed countries.

Distribution of supercomputers in the TOP500 list by country (November 2018)
Country Systems
 China
229
 United States
108
 Japan
31
 United Kingdom
20
 France
18
 Germany
17
 Ireland
12
 Canada
8
 Italy
6
 Korea, South
6
 Netherlands
6
 Australia
5
 India
4
 Poland
4
 Sweden
4
 Russia
3
 Saudi Arabia
3
 Singapore
3
 South Africa
2
 Spain
2
  Switzerland
2
 Taiwan
2
 Brazil
1
 Czechia
1
 Finland
1

Systems ranked No. 1 since 1976[edit]

Number of systems[edit]

By number of systems as of November 2018[28]:

New developments in supercomputing[edit]

In November 2014, it was announced that the United States was developing two new supercomputers to exceed China's Tianhe-2 in its place as world's fastest supercomputer. The two computers, Sierra and Summit, will each exceed Tianhe-2's 55 peak petaflops. Summit, the more powerful of the two, will deliver 150–300 peak petaflops.[29] On 10 April 2015, US government agencies banned selling chips, from Nvidia, to supercomputing centers in China as "acting contrary to the national security... interests of the United States";[30] and Intel Corporation from providing Xeon chips to China due to their use, according to the US, in researching nuclear weapons – research to which US export control law bans US companies from contributing – "The Department of Commerce refused, saying it was concerned about nuclear research being done with the machine."[31]

On 29 July 2015, President Obama signed an executive order creating a National Strategic Computing Initiative calling for the accelerated development of an exascale (1000 petaflop) system and funding research into post-semiconductor computing.[32]

In June 2016, Japanese firm Fujitsu announced at the International Supercomputing Conference that its future exascale supercomputer will feature processors of its own design that implement the ARMv8 architecture. The Flagship2020 program, by Fujitsu for RIKEN plans to break the exaflops barrier by 2020 (and "it looks like China and France have a chance to do so and that the United States is content – for the moment at least – to wait until 2023 to break through the exaflops barrier."[33]) These processors will also implement extensions to the ARMv8 architecture equivalent to HPC-ACE2 that Fujitsu is developing with ARM Holdings.[33]

Inspur has been one of the largest HPC system manufacturer based out of Jinan, China. As of May 2017, Inspur has become the third manufacturer to have manufactured 64-way system – a record which has been previously mastered by IBM and HP. The company has registered over $10B in revenues and have successfully provided a number of HPC systems to countries outside China such as Sudan, Zimbabwe, Saudi Arabia, Venezuela. Inspur was also a major technology partner behind both the supercomputers from China, namely Tianhe-2 and Taihu which lead the top 2 positions of Top500 supercomputer list up to November 2017. Inspur and Supermicro released a few platforms aimed at HPC using GPU such as SR-AI and AGX-2 in May 2017.[34]

Large machines not on the list[edit]

Some major systems are not listed on the list. The largest example is the NCSA's Blue Waters which publicly announced the decision not to participate in the list[35] because they do not feel it accurately indicates the ability for any system to be able to do useful work.[36] Other organizations decide not to list systems for security and/or commercial competitiveness reasons. Additional purpose-built machines that are not capable or do not run the benchmark were not included, such as RIKEN MDGRAPE-3 and MDGRAPE-4.

Computers and architectures that have dropped off the list[edit]

IBM Roadrunner[37] is no longer on the list (nor is any other using the Cell coprocessor, or PowerXCell).

Although Itanium-based systems reached second rank in 2004,[38][39] none now remain.

Similarly (non-SIMD-style) vector processors (NEC-based such as the Earth simulator that was fastest in 2002[40]) have also fallen off the list. Also the Sun Starfire computers that occupied many spots in the past now no longer appear.

The last non-Linux computers on the list – the two AIX ones – running on POWER7 (in July 2017 ranked 494th and 495th[41] originally 86th and 85th), dropped off the list in November 2017.

See also[edit]

References[edit]

  1. ^ A. Petitet, R. C. Whaley, J. Dongarra, A. Cleary (24 February 2016). "HPL – A Portable Implementation of the High-Performance Linpack Benchmark for Distributed-Memory Computers". ICL – UTK Computer Science Department. Retrieved 22 September 2016.
  2. ^ "An Interview with Jack Dongarra by Alan Beck, editor in chief HPCwire". Archived from the original on 28 September 2007.
  3. ^ a b "Statistics on Manufacturers and Continents".
  4. ^ "The TOP25 Supercomputer Sites". Archived from the original on 23 January 2016. Retrieved 4 January 2015.
  5. ^ "Where does Asia stand? This rising supercomputing power is reaching for real-world HPC leadership". Retrieved 4 January 2015.
  6. ^ Rpeak – This is the theoretical peak performance of the system. Measured in PFLOPS.
  7. ^ "Sublist Generator". Retrieved 26 June 2018.
  8. ^ "HPCG - June 2018 | TOP500 Supercomputer Sites". www.top500.org. Retrieved 2018-07-02.
  9. ^ "Top500 – List Statistics". Top500.org. November 2017.
  10. ^ "Linux Runs All of the World's Fastest Supercomputers". The Linux Foundation. Retrieved November 30, 2017.
  11. ^ "Microsoft Windows Azure".
  12. ^ "Magic Cube – Dawning 5000A, QC Opteron 1.9 GHz, Infiniband, Windows HPC 2008".
  13. ^ "ORIGIN 2000 195/250 MHz | TOP500 Supercomputer Sites". www.top500.org. Retrieved 2017-11-15.
  14. ^ "PEZY-SC2 - PEZY". Retrieved 2017-11-17.
  15. ^ "The 2,048-core PEZY-SC2 sets a Green500 record". WikiChip Fuse. 2017-11-01. Retrieved 2017-11-15. Powering the ZettaScaler-2.2 is the PEZY-SC2. The SC2 is a second-generation chip featuring twice as many cores – i.e., 2,048 cores with 8-way SMT for a total of 16,384 threads. [..] The SC incorporated two ARM926 cores and while that was sufficient for basic management and debugging its processing power was inadequate for much more. The SC2 uses a hexa-core P-Class P6600 MIPS processor which share the same memory address as the PEZY cores, improving performance and reducing data transfer overhead. With the powerful MIPS management cores, it is now also possible to entirely eliminate the Xeon host processor. However, PEZY has not done so yet.
  16. ^ "TOP500 Supercomputer Sites". November 2018.
  17. ^ a b "China Tops Supercomputer Rankings with New 93-Petaflop Machine – TOP500 Supercomputer Sites".
  18. ^ ABCI. "ABCI". AI Bridging Cloud Infrastructure. National Institute of Advanced Industrial Science and Technology (AIST). Retrieved 25 June 2018.
  19. ^ SuperMUC-NG. "SuperMUC-NG". Leibniz Supercomputing Centre. Leibniz-Rechenzentrum. Retrieved 14 November 2018.
  20. ^ "TOP500 Supercomputer Sites". November 2018.
  21. ^ "TOP500 Supercomputer Sites". June 2018.
  22. ^ "TOP500 Supercomputer Sites". November 2017.
  23. ^ "TOP500 Supercomputer Sites". June 2017.
  24. ^ "TOP500 Supercomputer Sites". June 2016.
  25. ^ "TOP500 Supercomputer Sites". November 2015.
  26. ^ "TOP500 Supercomputer Sites". June 2015.
  27. ^ "TOP500 Supercomputer Sites". November 2014.
  28. ^ "List Statistics". Retrieved 14 November 2018.
  29. ^ Balthasar, Felix. "US Government Funds $425 million to build two new Supercomputers". News Maine. Archived from the original on 19 November 2014. Retrieved 16 November 2014.
  30. ^ "Nuclear worries stop Intel from selling chips to Chinese supercomputers". CNN. 2015-04-10. Retrieved 2016-08-17.
  31. ^ "US nuclear fears block Intel China supercomputer update".
  32. ^ Executive Order -- Creating a National Strategic Computing Initiative (Executive order), The White House – Office of the Press Secretary, 29 July 2015
  33. ^ a b Morgan, Timothy Prickett. "Inside Japan's Future Exascale ARM Supecomputer". The Next Platform. Retrieved 13 July 2016.
  34. ^ "Supermicro, Inspur, Boston Limited Unveil High-Density GPU Servers | TOP500 Supercomputer Sites". www.top500.org. Retrieved 2017-06-13.
  35. ^ Blue Waters Opts Out of TOP500 (article), 16 November 2012
  36. ^ Kramer, William, Top500 versus Sustained Performance – Or the Ten Problems with the TOP500 List – And What to Do About Them. 21st International Conference On Parallel Architectures And Compilation Techniques (PACT12), 19–23 September 2012, Minneapolis, MN, US
  37. ^ "Roadrunner – BladeCenter QS22/LS21 Cluster, PowerXCell 8i 3.2 GHz / Opteron DC 1.8 GHz, Voltaire Infiniband". Retrieved 4 January 2015.
  38. ^ "Thunder – Intel Itanium2 Tiger4 1.4GHz – Quadrics". Retrieved 4 January 2015.
  39. ^ "Columbia – SGI Altix 1.5/1.6/1.66 GHz, Voltaire Infiniband". Retrieved 4 January 2015.
  40. ^ "Japan Agency for Marine -Earth Science and Technology". Retrieved 4 January 2015.
  41. ^ "IBM Flex System p460, POWER7 8C 3.550GHz, Infiniband QDR". TOP500 Supercomputer Sites.

External links[edit]