In this era of mobile devices, everything seemed to have gone bite size now. There are smartphones, tablets, and fortified laptops for the masses’ disposal. Nonetheless, these dainty pieces of technology can only perform mundane tasks; the world needs supercomputers for more complex jobs.
Dating back to the 1960's, supercomputers branched off the personal computer in terms of functionality. Gradually, scientists in various fields have come to notice the viability of these humongous contraptions for weather forecasting, 3D nuclear test simulations, and molecular dynamics simulations, to mention a few.
As the quantitative world celebrates the arrival of the new “Tianhe-2” supercomputer, we run down some of history’s fastest.
9 Tianhe-1A (China)
Translated from Mandarin as “Milky Way Number One”, Tianhe-1A is one of the few petascale supercomputers in the world, and is capable of an Rmax (maximum range) of 2.6 petaflops. It uses a Linux based operating system, whereas it has Intel Xeon CPUs and Nvidia GPUs across its 183,368 processing cores. It’s currently located at the National Supercomputing Center in Tianjin, China, where it carries out computations for petroleum exploration and aircraft design.
8 SuperMUC (Germany)
The SuperMUC is the name of the Leibniz-Rechenzentrum’s (Leibniz Supercomputing Centre) newest supercomputer in Garching, near Munich, Germany. It employs IBM iDataPlex servers, 300TB of RAM, and an InfiniBand interconnect in order for its 147,456 cores to achieve a speed of 2.9 petaflops. To compensate for its tremendous energy usage, costs are cut by literally cooling chips and memory with 104 degrees Fahrenheit of freezing water.
7 Vulcan (U.S.A.)
Located at the Lawrence Livermore National Laboratory (LLNL) in California, the Vulcan supercomputer is devoted to both government use and collaborative endeavours of industry and research universities. Its 393,216 cores run at 4.3 petaflops, which are based on IBM’s Blue Gene/Q supercomputing technology. It also consists of 24 racks and 24,576 compute nodes, placing it within LLNL’s high performance facility, together with the Sequoia.
6 Juqueen (Germany)
The Juqueen supercomputer has recently replaced the ‘Jugene’ at the Forschungszentrum Jülich (Jülich Supercomputing Centre) as the institution’s lead machine. Boasting a 5-petaflop speed moulded off the Blue Gene/Q system, it is expected to perform complicated calculations in the areas of neuroscience, computational biology, energy, climate research, and in quantum physics. Its 458,752 cores only consume 2,301 kilowatts of power, also making it one of the most energy efficient supercomputers in the world.
5 Stampede (U.S.A.)
Installed at the University of Texas’ Advanced Computing Center, the Stampede supercomputer’s Dell PowerEdge servers are powered by Xeon processors and an InfiniBand interconnect, which gives it its 5.2 petaflop speed. The base system has already been accepted by the U.S. National Science Foundation (NSF), successfully accomplishing 450,000 computational jobs that include seismic hazard mapping, ice sheet modeling for studying climate change, improving the imaging quality of brain tumors, and carbon dioxide capture.
4 Mira (U.S.A.)
Also called the ‘IBM Mira’, Mira is a petascale Blue Gene/Q supercomputer operated by the U.S. Department of Energy’s Agonne National Laboratory. Its 786,432 cores currently clock in at 8.6 petaflops, which consumes 3.9 megawatts of power. It’s also partially funded by the NSF, where it will be used for research in the fields of material science, climatology, seismology, and computational chemistry. Made as America’s answer to the Chinese Tianhe-1A, the exact cost of building it hasn’t been released, though it’s known that the laboratory used its U.S. $180 million grant as partial payment.
4. K Computer (Japan)
The K Computer is Fujitsu’s “super” creation for the RIKEN Advanced Institute for Computational Science campus in Kobe, Japan. It’s based on a distributed memory architecture composed of over 80,000 computer nodes, serving 10.5 petaflops, 705,024 Sparc cores, and a six dimensional torus interconnect called Tofu. This Linux kernel based machine’s name “kei” means “10 quadrillion”, reflecting the number of possibilities it can do. These include solving energy, sustainability, healthcare, climate change, industrial, and space challenges facing the Japanese society today.
3 Sequoia (U.S.A.)
Together with the Vulcan supercomputer, the Blue Gene/Q patterned IBM Sequoia supercomputer is one of the LLNL’s crown jewels. It was constructed by IBM for the National Nuclear Security Administration as part of its Advanced Simulation and Computing Program, as fully deployed on June 2012. With hit speeds of almost 17.2 petaflops, it has 1.6 million cores devoted to functions such as conducting nuclear weapon simulations, solving fluid dynamics problems, and further purposes such as astronomy, energy, study of the human genome, and climate change.
2 Titan (U.S.A.)
The Titan supercomputer, developed by Cray, is another machine funded by the U.S. Department of Energy, this time for the Oak Ridge National Laboratory in Oak Ridge, Tennessee. An upgrade from the facility’s previous processor, the Jaguar, it runs at a speed of 17.6 petaflops using AMD based Cray CPU's and Nvidia GPU's in its 560,640 cores. It was partially made available to researchers in early 2013, serving molecular scale physics, climate models, as well as applications in laser fusion, medical imaging, and energy combustion, to name some. The Titan only uses 8,209 kilowatts of power, making it another energy efficient supercomputer.
1 Tianhe-2 (China)
China follows up its Tianhe-1A with Tianhe-2, the new fastest supercomputer in the world that was delivered ahead of its two-year expected schedule. It boasts a record performance of 33.86 petaflops, using Ivy Bridge based Intel Xeons and Intel Xeon Phi for a total of 3.12 million cores. It consumes 17,808 kilowatts of power, which it uses for simulation, analysis, and government security applications for the National University of Defense Technology. Theoretically, it could even reach speeds of up to 54.9 petaflops, illustrating just how far Chinese technology has reached. After testing, the supercomputer will be officially housed at the National Supercomputing Center in Guangzhou, China.
Seeing these supercomputers do wonders for the advancement of the academic and the scientific communities is a good indication that technology hasn’t strayed too far its real purpose; to improve people’s lives.
In today’s situation where every new invention looks as if it propagates unwanted commercialism and vanity, hopefully, the supercomputers and people behind them will eventually get the recognition they deserve.