Reportedly the world's fastest supercomputer, Perlmutter, has been powered up at the National Energy Research Scientific Computing Center (NERSC). Packing 6159 NVIDIA A100 Tensor Core GPUs, one of the supercomputer's first projects is to build the largest 3D map of the Universe.
Perlmutter Will Process Data From the Dark Energy Spectroscopic Instrument (DESI)
To make the world's largest 3D map of the Universe, the Perlmutter will use data from the Dark Energy Spectroscopic (DESI). DESI is an instrument used to capture pictures of deep space and can capture more than 5,000 galaxies in one picture.
In a post on the NVIDIA blog, the company stated that using the power of NVIDIA's A100 Tensor Core GPUs, researchers can process the images from on night so that they know where to point DESI the next time.
Perlmutter is significantly faster than earlier systems:
Peparing a year’s worth of the data for publication would take weeks or months on prior systems, but Perlmutter should help them accomplish the task in as little as a few days.
The primary motive behind the 3D map is to "shed light on dark energy," which is said to be behind the continuous expansion of the universe. The supercomputer has been named after Saul Perlmutter, a scientist who won the 2011 Nobel Prize for his research on dark energy.
DYK #Perlmutter will quadruple the computational power currently @NERSC?
— NERSC (@NERSC) May 28, 2021
The @HPE_Cray EX supercomputer—containing @nvidia A100 Tensor Core GPUs + 3rd Gen @AMD EPYC CPUs— will be among the world's fastest for simulation, analysis, and #AI.
More: https://t.co/XeAcwkqIZJ | #HPC pic.twitter.com/b3hypWd7XI
Scientists also plan to use Perlmutter's computing prowess to create better biofuels and batteries by analyzing atomic interactions. According to NVIDIA, traditional supercomputers are "barely able" to process the math required to create simulations of atoms over a few nanoseconds.
But with NVIDIA's new supercomputer, scientists are planning to do just that.
Software Is Essential in Perlmutter
While Perlmutter packs cutting-edge hardware, the software is an integral part of the supercomputer as well.
NVIDIA has been focusing on AI for a long time and with Perlmutter, the company believes it is right on time for "AI supercomputing." Scientists believe that AI will play an important role in scientific breakthroughs.
CEO of NVIDIA, Jensen Huang, said:
Perlmutter’s ability to fuse AI and high performance computing will lead to breakthroughs in a broad range of fields from materials science and quantum physics to climate projections, biological research and more.
The blog post states that Perlmutter supports "OpenMP and other popular programming models in the NVIDIA HPC SDK the system uses."
This, in combination with RAPIDS—an open source code on data science for GPUs, will significantly increase the speed of scientific discovery using data.
