All supercomputers are super fast. Some are super faster than others. Find out the speediest ones in this list.
Before you enter into this list of the top Supercomputers, you must know what's a petaFLOPS — a quadrillion FLOPS. A measure of a supercomputer's performance, FLOPS or floating-point operations per second signifies the power of the machine in terms of its processing speed. Now, here are the fastest supercomputers in the world:
Fugaku
Named after an alternative moniker for Mount Fuji, Fugaku, developed jointly by RIKEN and Fujitsu, registered an initial Rmax performance of 416 petaFLOPS in the FP64 high-performance LINPACK benchmark. Built with the Fujitsu A64FX microprocessor, the ARM version 8.2A processor architecture provides the platform for its CPU. Its supercomputers adopt the Scalable Vector Extension. This monster's design is made for AI applications such as deep learning. It's used by RIKEN Center for Computational Science (R-CCS), Kobe for high-priority social and scientific issues.
Summit
Installed at Oak Ridge National Laboratory, this supercomputer, developed by IBM, can do 200 petaFLOPS. It's clocked 148.6 in the LINPACK benchmark. Summit is employed to provide solutions hitherto considered impossible for AI, human health, energy, and other research areas.
Sierra
Crafted by the Lawrence Livermore National Laboratory, it's harnessed by the nuclear scientists of NNSA — National Nuclear Security Administration, to carry out investigations on the stockpiles through simulation in lieu of underground blast tests.
Fitted with Power 9 processors and Volta graphics processing units, its peak capability is 125 petaFLOPs, during which it consumes only approximately 11 MW.
HPC5
The parallel computing units of High-Performance Computing — layer 5 can peak at 51.7 petaFlops. Located at Eni’s Green Data Center, it's combined with HPC4 to pull out 70 petaFlops.
Consisting of 1820 EMC PowerEdge C4140 nodes, each with 2 Gold 6252 24-core CPUs and 4 V100 GPU accelerators, it's the most powerful industrial supercomputer in the world.
Tianhe-2
Nearly 3,120,000 cores processors including Xeon E5 processors, custom processors and Xeon Phi coprocessors, helps Tianhe-2 deliver 33.9 petaFLOPS.
Located at National Super Computer Center in Guangzhou it's being used for education and research.
Marconi-100
Offering close to 32 theoretical peak petaFLOPS, this beast is fueled by POWER9 technology and V100 Tensor Core GPUs to make it the world’s smartest AI supercomputer.
Marconi100 supports European researchers, through the PRACE initiative, and Italian researchers through the Italian Supercomputing Resource Allocation initiative (ISCRA). It's being utilised for answering socio-economic challenges such as climate change, renewable energy, sustainable economics, and precision medicine.
