American institute supplies 6000 Nvidia A100 GPUs to supercomputer – computer



The Scientific Computing Center for US National Energy Research is commissioning Perlmutter, a supercomputer focused on artificial intelligence. The supercomputer contains more than six thousand Nvidia A100 accelerators and 1,500 AMD EPYC 7763 processors.

Perlmutter is placed in the Shyh Wang Hall facility in Nersc at the Berkeley Lab and named after Saul Perlmutter, a physicist who co-won the Nobel Prize in 2011 for his research into the speed at which the universe is expanding. More than 7,000 scientists can use the high-performance system, three to four times more powerful than the Nersc Cori supercomputer, which was the institute’s most powerful machine to date.

The block is based on the Shasta Van HPE Cray platform It is built in three phases. The first stage is ready, offering 1536 nodes, one for each AMD EPYC 7763 processor and four A100 GPUs are connected via NVlink. The system also provides a Luster file system with a capacity of 35 petabytes of flash memory. This gives scientists access to more than four exaflops of mixed precision computing power.

They can use computing power, among other things, to make a three-dimensional representation of the universe visible. Describes Nvidia. This requires massive amounts of data from an addressed dark energy spectrophotometer, which took weeks to months on the previous system, which should shrink Perlmutter for a few days.

Later this year, the second-stage supercomputer is expected to be upgraded with 3,072 processor nodes, each with an AMD EPYC 7763 processor and 512 GB of memory per node.


Leave A Reply

Your email address will not be published.