Second-place Frontier on the Oak Ridge Nationwide Laboratory clocked an HPL rating of 1.353 EFlop/s, which is similar as in June. Primarily based on the HPE Cray EX235z architectures and geared up with AMD third-generation EPYC 64C 2GHz processors, the Frontier techniques has 9,066,176 complete cores and likewise depends on HPE Slingshot interconnect for information switch.
The third U.S. system, Aurora, based mostly on the Argonne Management Computing Facility, was submitted with 1.012 EFlop/s on the HPL benchmark, which is unchanged from June. Constructed by Intel and based mostly on the HPE Cray EX-Intel Exascale Compute Blade, the Aurora system communicates via HPE Slingshot interconnect and makes use of Intel Xeon CPU Max Collection processors.
In fourth place on the record is JUPITER (JU Pioneer for Progressive and Transformative Exascale Analysis) on the EuroHPC/Jülich Supercomputing Centre in Germany. It was being commissioned on the time of the final itemizing and made a preliminary HPL worth of 793.4 Petaflop/s on a partial system. Now totally deployed, it has achieved a benchmark of precisely 1 Exaflop efficiency. It’s constructed on an Eviden’s BullSequana XH3000 direct liquid-cooled structure utilizing Nvidia Grace Hopper GH200 chips with a complete of 4,801,344 cores.
