Computers have become ubiquitous in every aspect of our life. They are used for everything from simple tasks like controlling streetlights and powering calculators to extremely complex tasks like charting orbits for space travel. The most complex tasks are performed by extremely powerful computers known as ‘Supercomputers.’ These supercomputers can perform quadrillions of floating-point calculations at the same time. Most supercomputers are actually a combination of a large number of computers that are capable of performing parallel processing.
Facts about Supercomputers
Supercomputers are currently being used to solve a host of problems. There is a very high chance that you have been exposed to the results given out by a supercomputer yourself. Let’s take a look into the fascinating world of super-fast calculations and computers the size of a giant hall.
1. Lenovo makes nearly 35% of all supercomputers in the world: Lenovo, known across the consumer PC market for their ultrabooks, are, in fact, the world’s largest manufacturers of supercomputers. As of November 2019, 34.9% of the top 500 fastest supercomputers were built and sold by Lenovo, putting it far ahead of its closest competitor, Inspur. Inspur holds 23.1% of the market share.
(Source: Business Wire)
2. China has almost half of the most powerful supercomputers in the world: Supercomputers require a lot of manpower to build and maintain, requiring personnel for a variety of tasks. They are also prohibitively expensive to install and run. So, it should come as no surprise that China, a country with both large amounts of manpower and a huge reserve of money, owns and operates nearly half of all the supercomputers in the world.
(Source: Top500.org)
3. Linux and Linux based OS is used in all of the top 500 fastest supercomputers: Windows may have the lion’s share of the consumer market when it comes to operating systems, but LINUX is the undisputed king of supercomputer OS. As of November 2019, all of the top 500 fastest supercomputers run on LINUX or LINUX based operating systems. The total share of LINUX across all supercomputers is estimated to be around 95%.
(Source: itfoss.com)
4. The USA has the top two fastest supercomputers: While China may be the pack leader when it comes to the sheer number of supercomputers at its disposal, the USA houses the two fastest supercomputers in the world. These computers, namely Summit and Sierra, boast of 1.5 times the speed of China’s TaihuLight but only consume 0.75 times the power.
(Source: The Verge)
5. The US government invested up to $430 million in supercomputing research: TheUSA’s position as the country with the most powerful supercomputers was achieved due to the exorbitant funding the government provides to create them. In 2017 the US government allocated $270 million in funding to the DOE, which funds supercomputer research. This was then backed with a commitment to raise 40% of the cost of the system in total, totaling an investment of $430 million over two years.
(Source: The Verge)
6. Japan has the most efficient supercomputers in the world: The Japanese people are known for their adherence to efficiency and this extends to their supercomputers as well. They have four of the top 10 most efficient systems in the world. Their A64FX has a peak efficiency of 16.8 GFLOPS/watt. For comparison, Summit has a peak efficiency of 14.71 GFLOPS/watt.
(Source: Green500)
7. The USA once blocked the export of Intel’s chips to China, leading China to develop the fastest supercomputer processor in the world: In 2105, China was on track to build the fastest supercomputer yet, the Sunway TaihuLight. This machine (currently active) was rumored to be developed to aid in nuclear tests simulations. In an effort to prevent it from being built, the US government banned an export license to Intel. Intel was, and still is, the leading producer of supercomputer processors. Undeterred, the Chinese government-funded the production of their own proprietary chips, which led to TaihuLight being built for nearly $30 million less than its allocated budget. It still stands as the most cost-efficient supercomputer in the world (cost for building to computing power)
(Source: BBC)
8. China will finish building the first exascale computer in 2020: Modern supercomputers are rated at Petaflops, which is one million billion floating-point operations performed in one second. The fastest supercomputer in the world, US’ Summit, is capable of operating at an astounding 148 Petaflops. However, according to the Chinese government reports that it will be ready to unveil the world’s first Exaflop capable computer. This computer will be capable of up to one billion billion calculations in one second, nearly five times the speed of Summit. The USA is not too far behind, claiming to be able to create a supercomputer capable of 1.5 exaflops capable machine by 2021.
(Source: nextplatform.com)
9. The supercomputer market was worth nearly $5.4 billion in 2019: Supercomputers are very expensive and useful machines. The market for these powerful machines is massive. In 2017 this market was worth $4.12 billion worldwide. It grew to $5.4 billion by the end of 2019 and is expected to grow at a combined annual growth rate of 9.5% for a value of $7.1 billion by the end of 2022. (Source: Market Study Report)
10. Cray – CDC 6600 was the first supercomputer and was as powerful as a scientific calculator: The first-ever computer to be classified as a supercomputer was Cray’s CDC – 6600. This qualification was bestowed upon it by virtue of it being nearly 100 times more powerful than the then consumer computers. It could function at a rate of 1000 flops. This would make it slightly less powerful than most modern scientific calculators, which can be used at a rate of around 1 MFLOP.
(Source: Computerhistory.org)
11. Supercomputers have been used to create AI that can beat human world champions at their own game: Artificial Intelligences are extremely potent at performing nearly any task. Programmers often use popular games like chess or Go to test their AI’s ability to learn complex tasks. One of the most daunting games for AIs to play is StarCraft 2. It is one of the most difficult games to master in the world, with the world champions having played it for over ten years. However, recently, Google’s AI, named AlphaStar, proved itself to be up to the task. AlphaStar was trained on Google’s Deepmind supercomputer, which allowed it to play millions of full-length games against itself to make itself better. Soon it was able to beat the reigning StarCraft 2 champion 5-0.
(Source: Deep Mind)
12. Human brains are entire orders of magnitudes more powerful than a supercomputer: As we have learned, supercomputers can only currently perform at around 150 petabytes. However, biologists have postulated that human brains can process and store information at a rate of 256 exabytes, which is about 300 billion times faster than the fastest supercomputers.
(Source: Wired UK)
13. The Washington Post’s supercomputer uses AI to write news stories: While the phrase ‘Robots will take over all the job’ is rather scary, there is always the security that there will always be some jobs that robots can’t take over. One of these jobs was thought to be journalism. However, this changed in 2017, when Heliograf, The Washington Post’s supercomputer trained AI, wrote the coverage of Iowa’s fourth congressional election in 2017. Since then, Heliograf has written several other stories that have been published in The Washington Post.
(Source: The Washington Post)
14. Spain’s second most powerful supercomputer lives right next to God: Spain’s second most powerful supercomputer resides in one of the most unlikely locations ever, in the basement of a grand cathedral! The Mare Nostrum, the supercomputer, resides in the Torre Giorna, a cathedral that was used for worship by Catholic Monks until the late 1960s before it was sold to the Spanish Government. It was then repurposed to house IBM’s Mare Nostrum in 2005, and the massive machine has resided there ever since.
(Source: Barcelona Supercomputing Center)
15. Parallella, the pocket-sized supercomputer, which costs $99: After reading so many amazing things about supercomputing, do you want to obtain and use one yourself? Well with Parallella, you can! Touted as the world’s smallest supercomputer, Parallella boats of an 18-core architecture, which can output up to 102 GFLOPS of computational prowess. This is paltry compared to the giants of the supercomputing world, but still a hefty amount o power, especially considering its price tag. The 16 core variant only costs $99!
(Source: Parallella.org)
Facts about Quantum Computers
Now that we have taken a deep look into supercomputers, we will take a look into the natural evolution of the same. This evolution is brought on due to the physical restriction of the solid-state transistors that supercomputer processors are based around.
These solid-state transistors have continuously been developed to be smaller, in accordance with Moore’s Law. The law dictates that the number of transistors that can fit in a processor will double every two years. However, these transistors have a hard limit for their size due to a phenomenon called quantum tunneling that allows electrons to simply pass through a barrier that is smaller than their wavelength. To circumnavigate this limit, researchers have been working on Quantum Computers. The first wave of this technology already exists in the form of quantum processors.
Quantum computers use quantum bits, called qubits, instead of binary bits to perform calculations. Binary bits can only hold two states, namely 1 and 0, and these states are completely independent of each other. Meanwhile, qubits can hold both states simultaneously. So, two qubits can be holding any one of four states at the same time. Hence depending on the factors that use these qubits, they can be used in two different calculations. This means that calculation performed by qubits are entire magnitudes of time quicker. Since qubits are quantum particles instead of electrons, they can also be much smaller than standard transistors. So, quantum computing is much faster and much smaller than a standard transistor-based computer.
Quantum computing is widely recognized to be the next major evolution of technology, joining the ranks of inventions like the steam engine and computers. Keeping this in mind, here is a look into the interesting facts and statistics related to Quantum Computing.
16. IBM’s quantum machines work with 16-qubits: IBM has been successful in creating a Q processor machine, which is the precursor to true quantum computers. Just like how classical computers with thousands of transistors were predated by ‘calculating machines’ which only used 10-20 transistors, the Q processor machines use only 16 qubits, while a true quantum computer would use thousands of these qubits in one processor.
(Source: IBM)
17. IBM’s Q machines are accessible to the public: IBM released their ‘Q – Experience’ in 2016, which was a cloud-based service where-in users can log in and use their quantum machines for whatever purpose they desire. They have also created their own programming language and compiler for the quantum machine to help in inducing users into the quantum realm.
(Source: IBM)
18. Most current research into creating compilers/languages for quantum machines are not done on quantum machines: While it is true that quantum machines exist, they are in a very rudimentary stage. Researchers can find themselves losing their progress because of decoherence of qubits. To prevent this instability from affecting the work of researchers, supercomputers are used to emulate qubits. However, even the most powerful supercomputers can only manage to simulate only 64 qubits.
(Source: phys.org)
19. Quantum computing might work due multiple parallel universes: Even though quantum machines have been created, scientists still aren’t 100% sure how exactly they work. One such explanation uses the theory of parallel universe. It states that qubits exist in multiple universes at the same time, and we observe them in all of these universes together until something forces them to collapse into one particular state, or universe. This is a far-fetched theory for sure, but it has gotten some support in the quantum community.
(Source: Sociable.co)
20. Cores of quantum machines work at close to -272.98o C, only 0.02o C above absolute zero: Qubits are very fragile and can lose their superimposed state very easily due to outside interference. One way to keep them stable is to keep them at very low temperatures. Google’s and IBM’s machines use this method for their cores. These cores are cooled using a special isotope of helium and lasers to achieve a temperature of 0.02 Kelvin, which is just 0.02o C above the absolute zero. This is about 150 times colder than deep space.
(Source: Intel)
21. China announced the creation of a $10 billion facility to research quantum computing: Countries across the globe have recognized the potential of quantum computing and have begun massive efforts to be able to create the first quantum computer. Leading this effort is China, whose government announced the creation of a $10 billion facility specifically to study quantum technology. They also awarded $1.2 billion in various grants to universities and labs for the same.
(Source: Wired UK)
22. The earliest consumer quantum computer will be available for purchase by 2050: Quantum computers are far out of the reach of the average consumer right now. IBM’s 16 qubit Q machines come with a heavy price tag of $15 million, and Google’s Sycamore is estimated to cost around $45 million. This price doesn’t even include the price of the super coolers needed by the cores to maintain temperatures of close to absolute zero.
(Source: Futurism.com)
23. Chinese scientists filed for 492 patents related to quantum technology by 2018: The massive efforts by the Chinese government to improve the research efforts into quantum technology have borne tangible fruits. Since 1980, Chinese scientists have filed 492 patents related to quantum technology, with 395 having been filed after 2016. This puts China far ahead of the rest of the world when it comes to unique patents being filed in this field. The second place is secured by the USA, which filed a measly 248 patents by 2018.
(Source: The Washington Post)
24. Google has managed to fit 53 qubits onto a processor: While the D wave machines available for public use only have 16 qubits available for use, Google’s ‘Sycamore’ boasts of 53 qubits working in tandem. Sycamore is the culmination of a 12-year long project undertaken by Google, firmly cementing them in the lead of quantum computer technology.
(Source: IEEE)
25. A problem which would have taken Summit 10,000 years to solve was solved in a mere 200 seconds by Google’s Sycamore: In one of Sycamore’s stress tests, Google gave it an impossible test to on which to work. This test would have taken the fastest supercomputer in the world, Summit, nearly 10,000 years to complete. Sycamore completed this in barely 200 seconds. That’s nearly 1.5 trillion times faster than Summit! What’s more, Sycamore isn’t even considered a true quantum computer, since it only houses 53 qubits. One can only imagine what an actual quantum computer will be capable of!
(Source: livescience.com)
26. Quantum computing will render modern encryption algorithms useless: Modern encryption uses a 256-bit key for encoding information. This is basically a 256 digit long number which is the multiplication of two unique prime numbers. Multiplication of any two numbers is very easy, but to figure out the two unique primes from the 256 digit number is practically impossible. In-fact modern encryption algorithms rely on the fact that it would take a classical computer more time than the Earth’s lifespan to solve this factorization. However, this is not a problem that quantum computers face. Since even Google’s Sycamore, not even a true quantum computer, is about 1.5 trillion times faster than the fastest supercomputer, prime number encryption is going to go out of use soon.
(Source: MIT Technology Review)
27. Quantum computers will use up to 1000 times less energy than their classical counterparts: Modern computers use electrons to perform all their calculations. These electrons need to be sourced from the power supply. This, coupled with the power needed for cooling giant racks of servers, makes supercomputers require a large amount of power to work. Summit requires nearly 25 kW/hr of power when operating at peak capacity. Meanwhile, Google’s Sycamore only needs 2034 W/hr of power.
(Source: Cornell University)
28. Quantum computers might be the breakthrough humanity needs to become an interstellar species: NASA plans on using the power of quantum computing to take humanity to the stars. They plan on using quantum computers to predict the motion of stars and asteroids and create ever more efficient routes for their spacecraft. This would make these spacecrafts infinitely more capable of making much longer trips, with much less fuel. Quantum computing can also be used for advanced material design to create hyper-efficient engines and fuels.
(Source: IAARC)
29. Microsoft’s quantum computing research is based upon a particle that most physicists don’t think even exist: As we have discussed, quantum technology is a field with continuously evolving research. Even then, there are some laws and techniques that are agreed upon by the community. However, Microsoft prefers to not listen to these ‘efficient’ and ‘real’ techniques. They are busy chasing after a particle that most physicists think doesn’t even exists. They have chosen to go this route because, on the off chance that the particle does exist, the qubits that it would represent would be nearly 100 times more stable than the one that the industry uses right now.
(Source: Scientific American)
30. Even your iPhone 34 will not be based on a quantum processor: While it is true that quantum computers are trillions of times faster than their classical counterparts, they will most likely not replace the humble home PC. They are only efficient when performing a large and complex task. This is because quantum computers can perform ‘n2’ tasks in ‘n’ seconds, and classical computers perform ‘n’ tasks at the same time. So, this implies that if a classical computer can get a task done in one second, it will be processed at the same speed as a quantum computer. So, since our computers can already perform up to 50 MFLOPS, quantum computers won’t be useful for everyday tasks. It’ll be like asking the heavyweight boxing champion to use all his strength to open a jar!
(Source: Business World)
31. The first use for a quantum computer might be just creating batteries: The first postulated use for quantum computing will possibly be chemical simulations. These simulations would bring around a new wave of material science, which would lead to the creation of new and more efficient batteries. Modern batteries are severely limited in their capacity to hold power by virtue of the materials used in their construction for holding and transferring ions. This could be changed by inventing new materials using quantum computing.
(Source: Wired)
32. TU Delft has a network to send information using rudimentary teleportation!: Quantum entanglement has often been hailed as the precursor to teleportation. Quantum entanglement involves two quantum particles that have been ‘entangled.’ These particles then can read and copy each other’s state across hundreds of kilometers, and this process is done instantly. TU Delft has demonstrated the power of this instant teleportation of information to create a small network of a quantum internet, that can transfer information anywhere in the network in less than a moment.
(Source: TU Delft)
33. Quantum computing can facilitate immortality: The human brain contains nearly 86 billion neurons. These neurons are all interconnected with each other to form the neural pathways that form the conscious of a human being. Even though artificial intelligence is making huge strides thanks to the power of supercomputers, it can never hope to match the human brain. While China and the USA are competing to be the first to create an exa-computer, which could theoretically match the processing speed of a brain, it could never be a true intelligence. This might change with the advent of the quantum computing. The power of quantum computing may allow researchers to create an actual sentient brain and could even allow for the human brain to be uploaded into an artificial shell that could theoretically live forever!
(Source: BBC)
34. Quantum computing could lead to the cure for cancer and Alzheimer’s: Modern medicine is designed using massive supercomputers to aid in complicated processes like protein folding and predicting the effects and side effects of various chemicals. Cancer research has been furthered a lot by the advent of these techniques. However, quantum computing might finally provide the processing power needed by researches to break the barrier to curing humanity’s most prolific killers.
(Source: BBC)
References and Data Sources:
- Business Wire
- Top500.org
- itfoss.com
- The Verge
- The Verge
- Green500
- BBC
- thenextplatform.com
- Market Study Report
- Computerhistory.org
- Deep Mind
- Wired UK
- The Washington Post
- Barcelona Supercomputing Center
- Parallella.org
- IBM
- IBM
- phys.org
- Sociable.co
- Intel
- Wired UK
- Futurism.com
- The Washington Post
- IEEE
- livescience.com
- MIT Technology Review
- Cornell University
- IAARC
- Scientific American
- Business World
- Wired
- TU Delft
- BBC
- BBC