Completed in partnership with Intel and Hewlett Packard Enterprise (HPE), the Aurora supercomputer will become the US’s second large-scale computer after the AMD-powered Frontier supercomputer when it goes live. Aurora will also surpass Frontier to become the world’s most powerful exascale-class computer.The Aurora supercomputer at Argonne National Laboratory will feature 63,744 Intel Data Center GPU Max series cards and 21,248 Intel Xeon CPU Max series processors. Aurora will also be the largest Xeon Max CPU-based system and the world’s largest GPU cluster. Aurora is expected to offer very high processing power for artificial intelligence, science and engineering.
What is Aurora
Developed in partnership with Intel, HPE, and DOE, the Aurora supercomputer is a system designed to unlock the potential of the three pillars of high-performance computing (HPC): simulations, data analytics and artificial intelligence (AI) at an enormous scale. This setup has more than 1,024 storage nodes providing 220 terabytes (TB) of capacity with a total bandwidth of 31TB. Later this year, Aurora will offer more than 2 exaflops of theoretical processing power when it launches.For comparison: The most powerful system available today is AMD-powered Frontier and offers 1,194 exaflops of processing power. Frontier is almost 2.5 times stronger than its second-ranked competitor. Intel-powered Aurora will be almost 2 times more powerful than Frontier.
Why are supercomputers like Aurora important?
From tackling climate change to curing deadly diseases, researchers face enormous challenges that require advanced computing technologies on a large scale. Aurora and similar supercomputers exist to meet the needs of the HPC and AI communities by providing the tools necessary to push the boundaries of scientific research. The supercomputer is optimized differently than traditional computers or servers. Thanks to target-oriented optimization, the most accurate results are obtained in the shortest time.
22817
