They should run at FP1 precision for maximum AI output quality. Sigh
Los Alamos Lab powers up Nvidia-laden Venado supercomputer
Los Alamos National Laboratory (LANL) has flipped the switch on its Venado supercomputer – a machine capable of bringing ten exaFLOPS of performance to bear on AI workloads for the Department of Energy. Announced at the ISC high performance computing conference in 2022, Venado is among the first supercomputers to be built …
COMMENTS
-
Tuesday 16th April 2024 11:04 GMT HuBo
A grace first?
I wonder if Venado is the first supercomputer to go live with Nvidia's Grace CPU in it? Last November's Top500 lists 10 systems with H100s in them (8 with Xeons, 2 with EPYCs) but I couldn't find any Grace-Grace nor Grace-Hopper system in that list (could be me mis-searching ...). If Venado is first-with-Grace then that's big news IMHO, and hopefully we do get a full set of HPL, HPCG, Green500, HPL-MxP, and Graph500 benchmarks from it (eg. to compare with A64FX, and EPYC/Xeon/POWER).