Intel finally announced the full specifications of the Aurora supercomputer; A device designed for the Argonne National Laboratory in the United States.
Intel’s Aurora supercomputer has been delayed for a long time, but its construction is finally being completed. With the support of Intel’s Xeon CPU Max and Xeon GPU Max series, this system has now reached an extraordinary processing power of two exaflops (Exaflop), while the initial goal for it was equivalent to one exaflop. Thus, Aurora is equal in terms of power to the current fastest supercomputer in the world, the Frontier equipped with AMD processors.
Intel announced that Aurora will have a total of 10,624 nodes, including 21,248 Xeon processors based on the Sapphire Rapids-SP family and 3,744 GPUs based on Ponte Vecchio. This very powerful supercomputer uses an impressive bandwidth of 2.12 petabytes per second and a maximum of 0.69 petabytes of dual bandwidth, which is an insane figure.