The US Department of Energy has unveiled its new supercomputer presented as the most powerful in the world. Delivering 200 petaflops, it is specifically the first of its kind to have accomplished an exascale scientific calculation.
In terms of supercomputers, the race for power that the great nations are engaged in is moving towards a specific goal: to reach an “exascale” computing capacity, that is to say a billion billion operations per second. In the United States, the Oak Ridge National Laboratory is taking an important step in this direction with its new supercomputer.
Called Summit, this power monster consisting of 4,608 IBM servers each containing two Power9 processors with 22 cores and six Nvidia Tesla V100 graphics processors, delivers a computing capacity of 200 petaflops, thus 200x 1015 floating point operations per second. This makes it the most powerful supercomputer in the world. Most importantly, Summit has produced the first ever exascale computation, a comparative genomics computation for health bioenergy research. According to the release, the maximum computation speed reached 1.88 exaops, or 1.88 x 1018 operations per second, but on whole numbers. The researchers say that the results of this calculation were identical to those made by Titan, the Oak Ridge supercomputer of 17.59 petaflops commissioned in 2012, which had worked much longer. For some operations, the statement said, researchers hope to climb to 3.3 exaops.
Thanks to this advance, the United States hope to have the first exaop supercomputer by 2021. In this quest, China has announced that it will launch the first exascale class supercomputer in 2020.
In the meantime, Summit will help advance astrophysical research, especially to study how supernovas create heavy elements such as gold and iron. The machine will also be used for atomic-scale simulations for the development of new materials (storage, conversion and energy production), the analysis of public health data to better understand the development of cancer in the population the United States. Finally, algorithms for automatic and deep learning will be used for genetic and biomedical analysis of human diseases.