Work in the most powerful systems. The most powerful supercomputers in the world

People still don’t fly to Mars, cancer hasn’t been cured yet, and we haven’t gotten rid of oil addiction. And yet there are areas where humanity has made incredible progress in recent decades. The computing power of computers is just one of them.

Twice a year, experts from the Lawrence Berkeley National Laboratory and the University of Tennessee publish the Top 500, which offers a list of the most powerful supercomputers in the world.

Looking ahead a little, we suggest you taste these numbers in advance: the productivity of the representatives of the top ten is measured in tens of quadrillions of flops. For comparison: ENIAC, the first computer in history, had a power of 500 flops; Today, the average personal computer has hundreds of gigaflops (billions of flops), the iPhone 6 has approximately 172 gigaflops, and the PS4 has 1.84 teraflops (trillion flops).

Armed with the latest Top 500 from November 2014, Naked Science decided to figure out what the 10 most powerful supercomputers in the world are, and what problems require such tremendous computing power to solve.

10. Cray CS-Storm

Location: USA
Performance: 3.57 petaflops
Theoretical maximum performance: 6.13 petaflops
Power: 1.4 MW

Like almost all modern supercomputers, including each of those presented in this article, CS-Storm consists of many processors united into a single computer network based on the principle of a massively parallel architecture. In reality, this system consists of many racks (“cabinets”) with electronics (nodes consisting of multi-core processors), which form entire corridors.

Cray CS-Storm is a whole series of supercomputer clusters, but one of them still stands out from the rest. Specifically, it is the mysterious CS-Storm, which is being used by the US government for unknown purposes and in an unknown location.

What is known is that American officials bought extremely efficient in terms of energy consumption (2386 megaflops per 1 Watt) CS-Storm with a total number of cores of almost 79 thousand from the American company Cray.

The manufacturer's website, however, says that CS-Storm clusters are suitable for high-performance computing in the fields of cybersecurity, geospatial intelligence, pattern recognition, seismic data processing, rendering and machine learning. Somewhere in this series, the use of the government CS-Storm probably settled down.

CRAY CS-STORM

9. Vulcan – Blue Gene/Q

Location: USA
Performance: 4.29 petaflops
Theoretical maximum performance: 5.03 petaflops
Power: 1.9 MW

“Vulcan” was developed by the American company IBM, belongs to the Blue Gene family and is located at the Lawrence Livermore National Laboratory. The supercomputer, owned by the US Department of Energy, consists of 24 racks. The cluster began operating in 2013.

Unlike the already mentioned CS-Storm, the scope of application of Vulcan is well known - various scientific research, including in the field of energy, such as modeling natural phenomena and analyzing large amounts of data.

Various scientific groups and companies can gain access to the supercomputer by submitting an application to the High Performance Computing Innovation Center (HPC Innovation Center), based at the same Livermore National Laboratory.

Supercomputer Vulcan

8. Juqueen – Blue Gene/Q

Location: Germany
Performance: 5 petaflops
Theoretical maximum performance: 5.87 petaflops
Power: 2.3 MW

Since its launch in 2012, Juqueen has been the second most powerful supercomputer in Europe and the first in Germany. Like Vulcan, this supercomputer cluster was developed by IBM as part of the Blue Gene project, and belongs to the same generation Q.

The supercomputer is located in one of the largest research centers in Europe in Jülich. It is used accordingly - for high-performance computing in various scientific research.

Juqueen supercomputer

7. Stampede – PowerEdge C8220

Location: USA
Performance: 5.16 petaflops
Theoretical maximum performance: 8.52 petaflops
Power: 4.5 MW

Located in Texas, Stampede is the only cluster in the top ten of the Top 500 that was developed by the American company Dell. The supercomputer consists of 160 racks.

This supercomputer is the most powerful in the world among those used exclusively for research purposes. Access to Stampede facilities is open to scientific groups. The cluster is used in a wide range of scientific fields - from precise tomography of the human brain and earthquake prediction to identifying patterns in music and language structures.

Supercomputer Stampede

6. Piz Daint – Cray XC30

Location: Switzerland
Performance: 6.27 petaflops
Theoretical maximum performance: 7.78 petaflops
Power: 2.3 MW

The Swiss National Supercomputing Center (CSCS) boasts the most powerful supercomputer in Europe. The Piz Daint, named after the Alpine mountain, was developed by Cray and belongs to the XC30 family, within which it is the most productive.

Piz Daint is used for various research purposes, such as computer simulations in the field of high energy physics.

Supercomputer Piz Daint

5. Mira – Blue Gene/Q

Location: USA
Performance: 8.56 petaflops
Theoretical maximum performance: 10.06 petaflops
Power: 3.9 MW

The Mira supercomputer was developed by IBM as part of the Blue Gene project in 2012. Argonne National Laboratory's High Performance Computing Division, which houses the cluster, was created with government funding. The rise in interest in supercomputing technology from Washington in the late 2000s and early 2010s is believed to be due to rivalry with China in this area.

Located on 48 racks, Mira is used for scientific purposes. For example, the supercomputer is used for climate and seismic modeling, which allows obtaining more accurate data on predicting earthquakes and climate change.

Supercomputer Mira

4. K Computer

Location: Japan
Performance: 10.51 petaflops
Theoretical maximum performance: 11.28 petaflops
Power: 12.6 MW

Developed by Fujitsu and located at the Institute of Physicochemical Research in Kobe, the K Computer is the only Japanese supercomputer to appear in the top ten of the Top 500.

At one time (June 2011), this cluster took first position in the ranking, becoming the most productive computer in the world for one year. And in November 2011, K Computer became the first in history to achieve power above 10 petaflops.

The supercomputer is used in a number of research tasks. For example, for forecasting natural disasters (which is important for Japan due to the increased seismic activity of the region and the high vulnerability of the country in the event of a tsunami) and computer modeling in the field of medicine.

Supercomputer K

3. Sequoia – Blue Gene/Q

Location: USA
Performance: 17.17 petaflops
Theoretical maximum performance: 20.13 petaflops
Power: 7.8 MW

The most powerful of the four supercomputers of the Blue Gene/Q family, which are in the top ten of the rating, is located in the United States at the Livermore National Laboratory. IBM developed Sequoia for the National Nuclear Security Administration (NNSA), which needed a high-performance computer for a very specific purpose: simulating nuclear explosions.

It is worth mentioning that real nuclear tests have been prohibited since 1963, and computer simulation is one of the most acceptable options for continuing research in this area.

However, the power of the supercomputer was used to solve other, much more noble problems. For example, the cluster managed to set performance records in cosmological modeling, as well as in creating an electrophysiological model of the human heart.

Sequoia supercomputer

2. Titan – Cray XK7

Location: USA
Performance: 17.59 petaflops
Theoretical maximum performance: 27.11 petaflops
Power: 8.2 MW

The most productive supercomputer ever created in the West, as well as the most powerful computer cluster under the Cray brand, is located in the United States at the Oak Ridge National Laboratory. Despite the fact that the supercomputer at the disposal of the US Department of Energy is officially available for any scientific research, in October 2012, when Titan was launched, the number of applications exceeded all limits.

Because of this, a special commission was convened at the Oak Ridge Laboratory, which selected only 6 of the most “advanced” projects out of 50 applications. Among them, for example, modeling the behavior of neutrons in the very heart of a nuclear reactor, as well as forecasting global climate changes for the next 1-5 years.

Despite its computing power and impressive dimensions (404 square meters), Titan did not last long on the pedestal. Just six months after the triumph in November 2012, American pride in the field of high-performance computing was unexpectedly supplanted by a native of the East, surpassing the previous leaders of the ranking in an unprecedented way.

Supercomputer Titan

1. Tianhe-2 / Milky Way-2

Location: China
Performance: 33.86 petaflops
Theoretical maximum performance: 54.9 petaflops
Power: 17.6 MW

Since its first launch, Tianhe-2, or Milky Way-2, has been the leader of the Top-500 for about two years. This monster is almost twice as powerful as the No. 2 in the ranking – the TITAN supercomputer.

Developed by the People's Liberation Army Defense Science and Technology University and Inspur, Tianhe-2 consists of 16 thousand nodes with a total number of cores of 3.12 million. The RAM of this colossal structure, which occupies 720 square meters, is 1.4 petabytes, and the storage device is 12.4 petabytes.

Milky Way 2 was designed at the initiative of the Chinese government, so it is not surprising that its unprecedented power appears to serve the needs of the state. It was officially stated that the supercomputer is engaged in various simulations, analyzing huge amounts of data, as well as ensuring the national security of China.

Considering the secrecy inherent in Chinese military projects, one can only guess what kind of use the Milky Way-2 receives from time to time in the hands of the Chinese army.

Supercomputer Tianhe-2

For the fifth time in a row, China's Tianhe-2 (Milky Way 2) has become the fastest supercomputer in the world with a performance of 33.86 petaflops or quadrillion floating point operations per second. This is the verdict of the TOP500 list of the most powerful supercomputers, which is released twice a year.

Despite the expected result, there is still some interesting information in the latest edition. The United States still has more systems on the list than any other country, with 233 vehicles (compared to 231 six months ago and 265 a year ago). Second and third places are occupied by systems from the USA, while 141 machines from the list are located in Europe. It is noteworthy that three of the new computers belong to the Chinese company Lenovo, although China itself is represented by only 37 supercomputers, compared to 61 last year.
The average performance of the TOP500 has increased significantly over the past 6 months. The total power of all 500 supercomputers was 363 petaflops/s, which is noticeably more than 309 last November and 274 a year ago. 98% of systems use processors with six cores or more, while at least 88.2% have 8 cores per processor. Eighty-eight of the five hundred systems used accelerators/coprocessors, including Nvidia (52), ATI Radeon (4), and Intel Xeon Phi (33). The four systems use a combination of Xeon and Nvidia processors.
The top 10 consists of machines launched in 2011 and 2012, with the exception of a new entrant from Saudi Arabia at number 7. Here's what the list of the 10 most powerful supercomputers in the world looks like.

Tianhe-2: TH-IVB-FEP Cluster; National Supercomputing Center in Guangzhou, China; 3.12 million cores (33.86 Pflops/s).
Titan: Cray XK7 System, Oak Ridge National Laboratory, USA. 560,640 cores (17.59 Pflops/s).
Sequoia: IBM BlueGene/Q System, Livermore National Laboratory 1.57 million cores, (17.2 Pflops/s).
K Computer: SPARC64 system with 705,024 cores at RIKEN Institute of Advanced Computing Science at the Institute of Physical and Chemical Research (RIKEN), Japan. (10.5 Pflops/s).
Mira: IBM BlueGene/Q; DOE/SC/Argonne National Laboratory, USA; 786,000 IBM cores. (8.59 Pflops/s).
Piz Daint: Cray XC30 with 116,000 cores from Xeon and Nvidia; located at the Swiss National Computing Centre. (6.27 Pflops/s).
Shaheen II: Cray XC40 System. King Abdullah University of Science and Technology in Saudi Arabia. (5.536 Pflops/s).
Stampede: Dell PowerEdge C8220 system with 462,462 Xeon Phi cores at the University of Texas (5.17 Pflops/s).
JUQUEEN: BlueGene/Q, 458,752 IBM cores. Jülich Research Center, Germany. (5 Pflops/s).
Vulcan: BlueGene/Q, 393,216 cores IBM, US Department of Energy.

It must be remembered that the situation can change dramatically if someone creates a real quantum computer. IBM has set the record straight by building a 50-qubit computer (with a current maximum of 4) that could be more powerful than any system on this list.
Meanwhile, the US Department of Energy has ordered two IBM/Nvidia systems in a $425 million deal. Delivery of the machines is scheduled for 2017 and 2018, and peak power could reach 150 petaflops.

The K Computer supercomputer, which previously occupied first place, has been pushed to third place. Its performance is 11.28 Pflops (see Figure 1). Let us recall that FLOPS (FLoating-point Operations Per Second, FLOPS) is a unit of measurement of computer performance, which shows how many floating point operations per second a given computing system is capable of performing.

K Computer is a joint development of the Rikagaku Kenkiyo Institute of Physical and Chemical Research (RIKEN) and Fujitsu. It was created as part of the High-Performance Computing Infrastructure initiative led by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT). The supercomputer is installed on the territory of the Institute of Advanced Computing Sciences in the Japanese city of Kobe.

The supercomputer is based on a distributed memory architecture. The system consists of more than 80,000 compute nodes and is housed in 864 racks, each of which accommodates 96 compute nodes and 6 I/O nodes. The nodes, each containing one processor and 16 GB of RAM, are interconnected in accordance with the “six-dimensional loop / torus” topology. The system uses a total of 88,128 eight-core SPARC64 VIIIfx processors (705,024 cores) manufactured by Fujitsu using 45 nm technology.

This general purpose supercomputer provides high levels of performance and support for a wide range of applications. The system is used to conduct research in the fields of climate change, disaster prevention and medicine.

The unique water cooling system reduces the likelihood of equipment failure and reduces overall energy consumption. Energy savings are achieved through the use of highly efficient equipment, a heat and electricity cogeneration system and an array of solar panels. In addition, the mechanism for reusing waste water from the cooler reduces the negative impact on the environment.

The building in which K Computer is located is earthquake-resistant and can withstand earthquakes of magnitude 6 or more on the Japanese scale (0-7). To more efficiently accommodate equipment racks and cables, the third floor, measuring 50 × 60 m, is completely free of load-bearing columns. Modern construction technologies have made it possible to ensure an acceptable load level (up to 1 t/m2) for the installation of racks, the weight of which can reach 1.5 tons.

SEQUOIA SUPERCOMPUTER

The Sequoia supercomputer installed at the Lawrence Livermore National Laboratory. Lawrence, has a performance of 16.32 Pflops and ranks second in the ranking (see Figure 2).

This petaflop supercomputer, developed by IBM based on Blue Gene/Q, was created for the US National Nuclear Security Administration (NNSA) as part of the Advanced Simulation and Computing program.

The system consists of 96 racks and 98,304 compute nodes (1024 nodes per rack). Each node includes a 16-core PowerPC A2 processor and 16 GB of DDR3 RAM. In total, 1,572,864 processor cores and 1.6 PB of memory are used. The nodes are connected to each other in accordance with the “five-dimensional torus” topology. The area occupied by the system is 280 m2. Total energy consumption is 7.9 MW.

The Sequoia supercomputer was the first in the world to carry out scientific calculations that required more than 10 Pflops of computing power. Thus, the HACC cosmology simulation system required about 14 Pflops when running in the 3.6 trillion particle mode, and when running the Cardiod project code for simulating the electrophysiology of the human heart, performance reached almost 12 Pflops.

TITAN SUPERCOMPUTER

The Titan supercomputer, installed at the Oak Ridge National Laboratory (ORNL) in the USA, was recognized as the world's fastest supercomputer. In Linpack benchmark tests, its performance was 17.59 Pflops.

Titan implements a hybrid CPU-GPU architecture (see Figure 3). The system consists of 18,688 nodes, each equipped with a 16-core AMD Opteron processor and an Nvidia Tesla K20X graphics accelerator. A total of 560,640 processors are used. Titan is an update to ORNL's previously used Jaguar supercomputer and occupies the same server cabinets (total area of 404 m2).

The ability to use existing power and cooling systems saved approximately $20 million during construction. The power consumption of the supercomputer is 8.2 MW, which is 1.2 MW more than the Jaguar, while its performance in floating point operations is almost 10 times higher.

Titan will primarily be used to conduct research in materials science and nuclear energy, as well as research related to improving the efficiency of internal combustion engines. In addition, it will be used to model climate change and analyze potential strategies to address its negative impacts.

THE "GREENEST" SUPERCOMPUTER

In addition to the Top500 rating, aimed at identifying the most high-performance system, there is a Green500 rating, which recognizes the “greenest” supercomputers. Here, the energy efficiency indicator (Mflops/W) is taken as a basis. At the moment (the latest release of the rating is November 2012), the leader of the Green500 is the supercomputer Beacon (253rd place in the Top500). Its energy efficiency indicator is 2499 Mflops/W.

Beacon is powered by Intel Xeon Phi 5110P coprocessors and Intel Xeon E5-2670 processors, so peak performance can reach 112,200 Gflops with a total power consumption of 44.9 kW. Xeon Phi 5110P coprocessors provide high performance with low power consumption. Each coprocessor has 1 teraflops of power (double precision) and supports up to 8 GB of GDDR5 memory with 320 Gbps bandwidth.

The Xeon Phi 5110P's passive cooling system is rated at 225W TDP, which is ideal for high-density servers.

SUPERCOMPUTER EURORA

However, in February 2013, reports emerged that the Eurora supercomputer, located in Bologna (Italy), surpassed Beacon in energy efficiency (3150 Mflops/watt versus 2499 Mflops/W).

Eurora is built by Eurotech and consists of 64 nodes, each of which includes two Intel Xeon E5-2687W processors, two Nvidia Tesla K20 GPU accelerators and other hardware. The dimensions of such a node do not exceed the dimensions of a laptop, but their performance is 30 times higher and power consumption is 15 times lower.

High energy efficiency in Eurora is achieved through the use of several technologies. Water cooling makes the greatest contribution. Thus, each supercomputer node is a kind of sandwich: central equipment at the bottom, a water heat exchanger in the middle, and another electronics unit at the top (see Figure 4).

Such high results are achieved thanks to the use of materials with good thermal conductivity, as well as an extensive network of cooling channels. When installing a new computing module, its channels are combined with the channels of the cooling system, which allows you to change the configuration of the supercomputer depending on specific needs. According to the manufacturers, the risk of leaks is eliminated.

The Eurora supercomputer elements are powered by 48-volt DC sources, the introduction of which has reduced the number of energy conversions. Finally, the warm water removed from computing equipment can be used for other purposes.

CONCLUSION

The supercomputer industry is actively developing and setting more and more new records for performance and energy efficiency. It should be noted that it is in this industry, like nowhere else, that liquid cooling and 3D modeling technologies are widely used today, since specialists are faced with the task of assembling a super-powerful computing system that would be able to function in a limited volume with minimal energy losses.

Yuri Khomutsky- Chief Project Engineer at I-Teco. He can be contacted at: [email protected]. The article uses materials from the Internet portal about data centers “www.AboutDC.ru - Solutions for Data Centers”.

Reading time: 7 min.

Until now, humanity has not reached the waste heaps of Mars, has not invented the elixir of youth, cars cannot yet soar above the ground, but there are several areas in which we have still succeeded. The creation of powerful supercomputers is just such an area. To evaluate the power of a computer, you need to determine which key parameter is responsible for this characteristic. This parameter is flops - a value that shows how many operations a PC can perform in one second. It is on the basis of this value that our Big Rating magazine ranked the most powerful computers in the world for 2017.

Supercomputer power - 8.1 Pflop/sec

This computer stores data that is responsible for the security of the United States military structure, and it is also responsible for the state of readiness for a nuclear attack, if necessary. Two years ago this machine was one of the most powerful and expensive in the world, but today Trinity has been replaced by newer devices. The system on which this supercomputer runs is Cray XC40, thanks to which the device can “perform” such a number of operations per second.

Mira

Supercomputer power – 8.6 Pflop/sec

Cray has released another supercomputer, Mira. The US Department of Energy ordered the production of this machine to coordinate its work. The area in which Mira operates is industry and the development of research potential. This supercomputer can calculate 8.6 petaflops per second.

Supercomputer power – 10.5 Pflop/sec

The name of this device immediately describes the power, the Japanese word “kei” (K) means ten quadrillion. This figure almost exactly describes its production capacity - 10.5 petaflops. The highlight of this supercomputer is its cooling system. Water cooling is used, which reduces the consumption of energy reserves and reduces the assembly speed.

Supercomputer power – 13.6 Pflop/sec

Fujitsu, a company from the Land of the Rising Sun, did not stop working, having released the K Computer supercomputer, they immediately began a new project. This project was the Oakforest-Pacs supercomputer, which is classified as a new generation of machines (Knights landing generation). Its development was commissioned by Tokyo and Tsukuba universities. According to the original plan, the device's memory was supposed to be 900 TB, and the performance of Oakforest-Pacs would be 25 quadrillion operations per second. But due to a lack of funding, many aspects were not finalized, so the power of the supercomputer was 13.6 petaflops per second.

Cori

Supercomputer power – 14 Pflop/sec

Just last year, Cori was in sixth place on the list of the most powerful supercomputers in the world, but with the crazy speed of technology development, it lost one position. This supercomputer is located in the United States, at the Lawrence Berkeley National Laboratory. Scientists from Switzerland, with the help of Cori, were able to develop a 45-qubit quantum computing machine. The production capacity of this supercomputer is 14 petaflops per second.

Supercomputer power – 17.2 Pflop/sec

Scientists from all over the world have long agreed that Sequoia is the fastest supercomputer on the planet. And this is not just so, because he is able to perform arithmetic calculations that would take 6.7 billion people 320 years, in one second. The size of the machine is truly amazing - it occupies more than 390 square meters and includes 96 racks. Sixteen thousand trillion operations or in other words 17.2 petaflops is the production capacity of this supercomputer.

Titan

Supercomputer power – 17.6 Pflop/sec

In addition to being one of the fastest supercomputers on the planet, it is also very energy efficient. The energy efficiency indicator is 2142.77 megaflops per watt of energy required for consumption. The reason for this low power consumption is the Nvidia accelerator, which provides up to 90% of the power needed for computing. In addition, the Nvidia accelerator has significantly reduced the area occupied by this supercomputer, now it only needs 404 square meters.

Supercomputer power – 19.6 Pflop/sec

The first launch of this device took place in 2013, in Switzerland, in the city of Lugano. Now the geolocation of this supercomputer is the Swiss National Supercomputing Center. Piz Daint is a combination of all the best features of the above machines, it has a very high energy efficiency and is very fast in calculations. Only one characteristic leaves much to be desired - the dimensions of this supercomputer; it occupies 28 huge racks. Piz Daint is capable of 19.6 petaflops of computing power per second.

Supercomputer power – 33.9 Pflop/sec

This device has the romantic name Tianhe, which in Chinese means “Milky Way”. Tianhe-2 was the fastest computer on the list of the 500 fastest and most powerful supercomputers. It can calculate 2507 arithmetic operations, which in petaflops is 33.9 Pflops/sec. The specialization in which this computer is used is construction; it calculates operations related to building and laying roads. Since its first launch in 2013, this computer has not lost its position in the lists, which proves that it is one of the best machines in the world.

Supercomputer power – 93 Pflop/sec

Sunway TaihuLight is the fastest supercomputer in the world, in addition to its enormous computing speed, it is also famous for its huge dimensions - it occupies an area of more than 1000 square meters. The 2016 international conference, which took place in Germany, recognized this supercomputer as the fastest in the world, and it still does not have a serious competitor in this regard. Its speed is three times higher than Tianhe-2, the closest supercomputer in this regard!

Technological progress does not stand still, it develops at cosmic speed, affects many aspects of human life, and has many both positive and negative sides. Technology of various types has now become available to humans: computers, robots and instruments. But the main purpose of any equipment is to simplify a person’s life; technology should not become meaningless entertainment that will only waste your time.

Many gamers know the company Maingear - it creates and sells powerful desktop computers. So this company, together with the famous company Razer, released the most powerful gaming computer in the world (for 2018) - R2 Razer Edition.

The new computer is primarily aimed at professional gamers. The PC has the latest generation processor from Intel, 64 GB of DDR4-2666 SDRAM RAM, several hard drives and SSDs, video cards from NVIDIA or AMD (to choose from) and a new water cooling system. This power is enough not only for the most demanding games in 4K resolution, but also for games that support virtual reality devices.

Sell the most powerful computer R2 Razer Edition There will be two configurations.

Basic model with AMD Ryzen 7 1800X with a frequency of 3.2 GHz, 8 GB of RAM, AMD Radeon R9 RX 580 video card and 1 TB hard drive. The computer costs $999 (65,000 rubles).
The most powerful model with Core i7-7700K, 16 GB of RAM, NVIDIA GeForce Titan X video card with 32 GB of memory and M.2 NVMe SSD with a capacity of 1 TB. The price of this configuration is $4,000 (260,000 rubles). But it can be made even more powerful by replacing the hard drive with an SSD, increasing the RAM, and so on. Then the approximate price of this computer will be 10,000 dollars (650,000 rubles).