In the last decade or so, demand for higher-efficiency computing systems have skyrocketed for market research, engineering & scientific work and associated businesses. This is where supercomputers come in the picture. A supercomputer is just like a computer but with a higher level of performance that is mostly measured in FLOPS (floating-point operations per second) instead of MIPS (million instructions per second). Supercomputers are helping us in pushing the boundaries of everything (from science to art) to levels that we humans could not reach on our own. The strength of these super giants has multiplied drastically in the last few decades. In 1985, Cray-2 was the most powerful supercomputer in the world and could process 1.9 billion FLOPS. However, the top 500 supercomputers available today are at least a million times more powerful. Given below is a list of ten of the most powerful supercomputers of 2020.
Trinity
With a power efficiency of 3.678 gigaflops/watt, Trinity is a Cray XC40 system managed and operated by Sandia National Laboratories and Los Alamos National Laboratory. The machine is able to peak out at nearly 20.2 teraflops. It inherited its names from America’s first nuclear test in 1945. The reason behind it is that Trinity is increasing the computing capabilities of NNSA Nuclear Security Enterprise. The supercomputer is being used to support the NNSA Stockpile Stewardship program’s certification and ensuring that the nuclear stockpile of the nation is safe and secure.
Piz Daint
Piz Daint is Europe’s most powerful computing system and ranks ninth on our list. It is named after an alpine mountain and is located in Lugano at the Swiss National Supercomputing Centre. It is an upgrade of a computing system that was built by Cray, an American company that was founded by Seymour Cray, the father of supercomputing. Its processors are developed by Intel and NVIDIA and have a speed of 21.2 petaflops. The supercomputer is involved in several research materials mainly in the field of physics, science, life sciences, data science, climatology and geophysics.
Frontera
Frontera is a new computing system that was built by Dell and is equipped by Intel. It was launched in September 2019 by the Texas Advanced Computing Centre (TAAC) at the University of Texas. It is the world’s fastest supercomputer that is located in a university and the eighth most powerful supercomputer in the world. Single its deployment in June, it has collaborated with almost 3 dozen scientific teams to research about quantum mechanics, climate models, black holes and drug design.
Summit
Until last year, Summit was the most powerful supercomputer in the world. It was built by IBM for Oak Ridge National Laboratory, Tennessee, U.S.A for their energy department. The supercomputer occupies the space of 2 basketball courts and has 2.41 million cores that helps it achieve 148.9 petaflops. From all the supercomputers in this list, Summit is the most energy efficient. It is mainly involved in civil scientific research and have already been a part of projects like simulation of earthquakes in urban territories, climate phenomena, the explosion of supernovae and many others.
Marconi-100
Marconi 100 is an Italian supercomputer that is one of the fastest supercomputers in the world right now. It was installed at Cineca in 2020 and is the largest supercomputer in Italy and Europe in the academic sector. Its processors were developed by NVIDIA and IBM and can peak at 32 petaflops. It will greatly help European and Italian researchers and equip them to face challenges like renewable energy, climate change and precision medicine.
Tianhe-2
Tianhe-2 is a supercomputer that is developed by the National University of Defense Technology, China. It processes at a great computing speed of 33.9 petaflops, which is double the performance of Sequoia or Titan. It is located at the National Super Computer Center, Guangzhou and is being used for education and research. Its processors were developed by Intel and Matrix and has a core count of nearly 3.12 million.
HPC5
High-Performance Computing – layer 5 or HPC5 is a set of computing units that has a peak performance of 51.7 petaflops. If you combine it with HPC4 (supercomputing system in operation from 2018), its processing power is 70 petaflops, which means that it can perform 70 million billion mathematical operations in just one second. Locates in Eni’s Green Data Centre, HPC5 is the most powerful industrial supercomputer on the planet.
Sunway TaihuLight
As of November 2018, Sunway TaihuLight is ranked the third most powerful supercomputer with a LINPACK rating of 93 petaflops. It is a Chinese supercomputer that uses a SW26010 processor. It is a chip designed by the Shanghai High-Performance IC Design Centre with more than 10.649 CPU cores in the entire system. For 2 years in this decade, it was the most powerful supercomputer in the world. The power efficiency of this supercomputing system us 6.051 gigaflops/watt.
Sierra
Sierra was built by IBM for the Lawrence Livermore National Laboratory located in California for NNSA. It has retained its second place in the TOP500 list for quite some time now. Its NVIDIA Volta GV100 accelerators and IBM Power9 processors have over 2.8 million cores and an unchanged performance at 94.6 petaflops. It is mainly dedicated to military research, especially the stimulation of nuclear weapons rather than underground testing.
Fugaku
Developed jointly by Fujitsu and RIKEN, supercomputer Fugaku was named after Mount Fuji (or an alternate name of the mountain). This supercomputer is built with a Fujitsu A64FX microprocessor and is installed at RCCS (RIKEN Centre for Computational Science) in Japan. It is designed especially for high priority scientific and social issues. Its best features include high computational performance, low power consumption of 30-40 MW, user convenience and a design that is suited for deep learning. It is almost 3 times faster than Summit with a peak of 415 petaflops. That is why, it is currently the most powerful supercomputer in the world.
