Photos: The world's 25 fastest supercomputers

1. Sunway TaihuLight

The Sunway TaihuLight supercomputer is almost three times faster than the next most powerful machine in the list.

Based in the National Supercomputing Center in the city of Wuxi, the Chinese system performs calculations to aid research and engineering work, ranging from climate modelling to advanced manufacturing.

Unlike most other supercomputers, the TaihuLight doesn't rely on Intel CPUs but instead utilises a custom ShenWei processor, a RISC CPU with 260 cores, and custom interconnects made in Wuxi.

Processor cores: 10,649,600

Max performance (Linpack benchmark): 93PFLOPS - quadrillion floating point operations per second

Memory: 1.3PB

Power consumption: 15,371kW

Image: National Supercomputing Center

2. Tianhe-2

Formerly the fastest supercomputer in the world, the Tianhe-2 is capable of more than 33 quadrillion calculations per second.

Otherwise known as the Milky Way 2, the Tianhe-2 supercomputer memory is based in the National Supercomputer Center in Guangzho, China.

The machine is capable of carrying out a massive number of operations in parallel, spreading tasks between its millions of cores. Each of the machine's 16,000 nodes has two Intel Xeon E5 Ivy Bridge processors and three Xeon Phi 31SP co-processors.

Processor cores: 3,120,000

Max performance (Linpack benchmark): 33.9PFLOPS - quadrillion floating point operations per second

Memory: 1PB

Power consumption: 17,808kW

Image: Zhao zilong - Imaginechina

3. Piz Daint

The fastest system in Europe and number three machine in the world is Piz Daint, based at the Swiss National Supercomputing Centre.

A recent upgrade to the Cray XC50 research machine doubled its performance, with Nvidia Tesla P100 GPUs added to its cluster of 2.2GHz Intel Xeon E5-2692 CPUs.

The Piz Daint has come a long way, entering the TOP500 supercomputer list at number 114 in 2012, but steadily climbing thanks to repeated upgrades.

Processor cores: 361,760

Max performance (Linpack benchmark): 19.6PFLOPS - quadrillion floating point operations per second

Memory: 340,480GB

Power consumption: 2,272kW

Image: Swiss National Supercomputing Center

4. Titan

Once the world's fastest, this US-based supercomputer slips down to fourth place.

Capable of handling more than 17 quadrillion calculations per second, Titan helps researchers at the US Oak Ridge National Laboratory probe climate change, alternate fuel sources, astrophysics and other major scientific challenges.

The Cray XK7 machine relies on a mix of processors, both AMD Opteron 6274 processors, clocked at 2.2GHz, and NVIDIA K20x accelerators.

Processor cores: 560,640

Max performance (Linpack benchmark): 17.6PFLOPS - quadrillion floating point operations per second

Memory: 710TB

Power consumption: 8,209kW

Image: Oak Ridge National Laboratory

5. Sequoia

Another former frontrunner, Sequoia was not only the most powerful supercomputer in its day but also one of the most efficient.

The IBM Blue Gene/Q system remains a processing powerhouse, and is used used by the US Lawrence Livermore National Laboratory to model massively complex processes, ranging from approximating the universe to the beating of a human heart. It relies on more than one million Power BQC cores.

Processor cores: 1,572,864

Max performance (Linpack benchmark): 17.2PFLOPS - quadrillion floating point operations per second

Memory: 1.6PB

Power consumption: 7,890kW

Image: Lawrence Livermore National Laboratory

6. Cori

The flagship machine at the National Energy Research Scientific Computing Center in Berkeley, California, Cori was a new entry on the list last year.

Named after the pioneering biochemist Gerty Cori, the system runs on a mix of Intel Xeon Haswell processors and Intel Xeon Phi many-core CPUs.

Processor cores: 622,336

Max performance (Linpack benchmark): 14PFLOPS - quadrillion floating point operations per second

Memory: 878,592GB

Power consumption: 3,939kW

Image: National Energy Research Scientific Computing Center

7. Oakforest-PACS

Japan's fastest supercomputer is based at Joint Center for Advanced High Performance Computing in Kashiwa.

Dropping down one place from last year, the Oakforest-PACS relies on an array of 68-core Intel Xeron Phi 7250 processors.

Processor cores: 556,104

Max performance (Linpack benchmark): 13.6PFLOPS - quadrillion floating point operations per second

Memory: 919,296GB

Power consumption: 2,719kW

Image: University of Tokyo

8. K computer

Japan's K computer has been a fixture in the top 10 for many years.

Situated in Kobe's RIKEN Advanced Institute for Computational Science, the K computer's ample processing power helps tackle pressing global challenges in areas ranging from disaster prevention to medical research.

It relies upon more than 700,000 SPARC64 VIIIfx processors, each rated at 2GHz.

Processor cores: 705,024

Max performance (Linpack benchmark): 10.5PFLOPS - quadrillion floating point operations per second

Memory: 1.4PB

Power consumption: 12,660kW

Image: Riken Advanced Institute for Computational Science

9. Mira

Capable of some 8.5 quadrillion calculations per second, Mira plays a pivotal role in research at the Argonne National Laboratory in Illinois.

The IBM BlueGene supercomputer helps US researchers model everything from the performance of jet engines to the inner workings of the human body.

Driving this machine are a host of 16-core, 1.6GHz Power BQC processors.

Processor cores: 786,432

Max performance (Linpack benchmark): 8.6PFLOPS - quadrillion floating point operations per second

Power consumption: 3,945kW

Image: Argonne National Laboratory

10. Trinity

The Trinity supercomputer helps the US test the effectiveness of its nuclear arsenal.

Based at Los Alamos National Laboratory, Trinity simulates nuclear explosions using thousands of 16-core Xeon E5-2698v3 processors.

Processor cores: 301,056

Max performance (Linpack benchmark): 8.1PFLOPS - quadrillion floating point operations per second

Power consumption: 4,233kW

Image: Los Alamos National Laboratory

11. United Kingdom Meteorological Office

It might not have a name, but the UKMO's supercomputer is the 11th fastest in the world. The Cray XC40 was first launched in 2016 and has cracked the top 25 for the first time in this list.

Processor cores: 89,856

Max performance (Linpack benchmark): 2,801.78 TFlop/s

Memory: unlisted

Power consumption: 1,347.84 kW

Image: UK Met Office

12. Stampede2

The Texas Advanced Computing Center's latest supercomputer is the Stampede2. It, like the original Stampede built in 2012, is a Dell machine. This is the first appearance of the Stampede2 on the list, which isn't surprising: It's a brand new machine.

Processor cores: 285,600

Max performance (Linpack benchmark): 6,807.5 TFlop/s

Memory: 403,200 GB

Power consumption: 1,890 kW

Image: Texas ACC

13. MareNostrum

Another newbie to the top 500 list is the Barcelona Supercomputing Center's MareNostrum.

Processor cores: 148,176

Max performance (Linpack benchmark): 6,227.2 TFlop/s

Memory: 331,776 GB

Power consumption: 1,380 kW

Image: Barcelona SCC

14. Marconi Intel Xeon Phi

It may have fallen two places since November 2016, but the CINECA Marconi is still an impressive machine.

Processor cores: 241,808

Max performance (Linpack benchmark): 6,223.04 TFlop/s

Memory: 455,168 GB

Power consumption: 1,600.20 kW

Image: CINECA

15. Pleiades

Venerable in supercomputer years, NASA's Pleiades has been around since 2011. Its place on the list has risen as high as number 7 and fallen as low as number 21, but the past few lists have seen it settle into the mid teens.

Processor cores: 241,108

Max performance (Linpack benchmark): 5,951.55 TFlop/s

Memory: 917,344 GB

Power consumption: 4,407.00 kW

Image: NASA

16. Theta

A few upgrades have helped Argonne's Theta supercomputer climb from 18 to 16 on this list.

Processor cores: 231,424

Max performance (Linpack benchmark): 5,884.62 TFlop/s

Memory: unlisted

Power consumption: 1,087 kW

Image Argonne National Lab

17. Hazel Hen

Höchstleistungsrechenzentrum Stuttgart is the home of this Cray supercomputer. It has been on the list since 2014.

Processor cores: 185,088

Max performance (Linpack benchmark): 5,640.17 TFlop/s

Memory: unlisted

Power consumption: 3,615 kW

Image: HLRS

18. Shaheen II

This Saudi Arabian supercomputer is slowly slipping down the ranks since its entry in 2015 at number 7. Now down to 18, it may be time for some hardware upgrades if King Abdullah University wants to stay competitive.

Processor cores: 196,608

Max performance (Linpack benchmark): 5,536.99 TFlop/s

Memory: unlisted

Power consumption: 2,834 kW

Image: King Abdullah Univ. of Science and Tech

19. Pangea

Pangea is the highest ranking machine from the private sector, making it a powerhouse--all its higher ranking brethren are government or university owned. Controlling company Total Exploration Production nearly doubled its capabilities in early 2016, taking it from number 32 back toward the top of the list.

Processor cores: 220,800

Max performance (Linpack benchmark): 5,283.11 TFlop/s

Memory: 54,000 GB

Power consumption: 4,150 kW

Image: Total Exploration Production

20. Stampede

This is the Texas Advanced Computing Center's second entry in the top 25--its other one (Stampede2) is at number 12. The original Stampede came online in 2012 and has been gradually upgraded since then.

Processor cores: 522,080

Max performance (Linpack benchmark): 5,168.11 TFlop/s

Memory: 204,800 GB

Power consumption: 4,510 kW

Image: Texas ACC

21. JUQUEEN

This German supercomputer has been around since 2012, when it occupied the number 8 spot.

Processor cores: 458,752

Max performance (Linpack benchmark): 5,008.86 TFlop/s

Memory: 458,752 GB

Power consumption: 2,301 kW

Image FZJ

22. Cheyenne

This relatively new machine came online in the latter half of 2016. It entered last November's list at number 20 and reappears at 22 this year.

Processor cores: 144,900

Max performance (Linpack benchmark): 4,788.19 TFlop/s

Memory: 202,752 GB

Power consumption: 1,727 kW

Image: UCAR/Marijke Unger

23. Vulcan

Lawrence Livermore National Laboratory brought Vulcan online in 2012. It was much smaller back then and only managed to make number 48 in the list. That was before upgrades brought it into the top 10, and it has been running the same setup since then (2013).

Processor cores: 393,216

Max performance (Linpack benchmark): 4,293.31 TFlop/s

Memory: 393,216 GB

Power consumption: 1,972 kW

Image: Lawrence Livermore National Lab

24. Abel

Another privately held machine, Abel does its work for Houston-based Petroleum Geo-Services.

Processor cores: 145,920

Max performance (Linpack benchmark): 4,042.46 TFlop/s

Memory: 583,680 GB

Power consumption: 1,800 kW

Image: Petroleum Geo-Services

25. European Center for Medium-Ranged Weather Forecasts

Another unnamed machine, the ECMWF's Cray XC40 is used to research medium-range weather forecasts. It has been online since mid 2016.

Processor cores: 126,468

Max performance (Linpack benchmark): 3,944.68 TFlop/s

Memory: unlisted

Power consumption: 1,897.02 kW

Image: ECMWF