US climate supercomputer to get $70m research boost – TechCentral.ie

The Department of Energy is looking to improve its complex climate model, which runs on the world’s fastest computer

Professional

Image: Dennis

The US Department of Energy (DoE) has announced seven new projects that will receive a total of $70 million in funding that will work to improve its climate prediction model.

The Energy Exascale Earth System Model (E3SM) is DoE’s advanced climate prediction model operating at exascale levels to provide extremely detailed simulations of weather systems, changes in ocean currents and decadal climate change.

The projects are expected to help develop more advanced simulations covering areas as specific as ice flow to provide a more accurate overall picture of climate change.

advertisement

E3SM runs on Frontier, the DoE’s supercomputer, which in May officially claimed the title of the world’s fastest computer. After exceeding 1 ExaFlop/s in the benchmark, Frontier has a theoretical maximum speed of 2 Exaflops, which can be expressed as two quintillion calculations per second. In total, the system includes over 9,400 CPUs and over 37,000 GPUs.

The funding was made available through the DoE Advanced Computing Funding Opportunity Announcement, with $70 million over the next five years and $14 million expected this fiscal year.

The list of winning projects includes Los Alamos National Laboratory (LANL) Antarctic System Science Framework in E3SMPacific Northwest National Laboratory (PNNL) Capturing mixed flood dynamics in E3SMand the University of Mexico Improved coupled climate simulations in E3SM through improved sea ice mechanics.

In June, the DoE Argonne National Laboratory announced the release of an updated version of E3SM, titled E3SMv2, which operates at more than twice the speed of the older model. It is capable of modeling global atmospheric conditions at a resolution of 100 km, as well as at a finer resolution of 25 km over North America.

Also included in the update was a new input-output (I/O) library, labeled SCORPIO, which uses more efficient algorithms and improved data caching to speed up the overall I/O process.

The funding comes shortly after the Inflation Reduction Act was signed into law by President Biden, which allocates $369 billion to renewable energy and energy security. This would provide companies with large tax incentives to adopt greener technologies, which could lead to the start-up of more sustainable businesses or drive green digital transformation.

“Being able to understand and predict what’s happening in a system as complex as planet Earth is critical to finding solutions to climate change,” said US Energy Secretary Jennifer M. Granholm.

© Denis Publishing House

Read more: climate change High performance computing HPCclaimte modeling on supercomputers

Leave a Reply

Your email address will not be published. Required fields are marked *