Analysis Despite growing alarm over rising power consumption in the data center, this spring's Green500 ranking of the world's most sustainable widely known supercomputers shows that the same power-hungry server accelerators behind the AI boom are also driving significant improvements in efficiency.
Nvidia GH200 AI Chip: What it can do!
Seven of the ten most power-efficient systems in the ranking used Nvidia GPUs, while half, including the top three, were powered by the Silicon Valley giant's Grace-Hopper Superchips (GH200).
We've covered that processor in depth before; at a high level, these franken chips combine a 72-core Grace CPU based on an Arm Neoverse V2 design with 480 GB of LPDDR5x memory and an H100 GPU with between 96 GB and 144 GB of HBM3 or HBM3e memory using a high-speed NvLink chip-to-chip connection.
According to Nvidia, the GH200 now makes up the bulk of Nvidia's HPC-centric installations and even claims a spot in the Top500's 10 most powerful widely known systems, with the debut of the Alps machine. In fact, the preAlps testbed took the number five spot in the Green500 ranking with an efficiency of 64.4 gigaFLOPS/watt. But, as we hinted before, this is far from the best the GH200 could handle.
