Benchmark
Teams choosing training and inference infrastructure for agentic AI workloads should weigh these MLPerf results, especially throughput per watt, when planning compute procurement and cost per token.
NVIDIA's Vera Rubin platform made its MLPerf debut, with NVIDIA sweeping every MLPerf Training v6.1 benchmark. In its first MLPerf Endpoints submission, the platform reported more than 10x generational gains in throughput per watt for agentic AI inference. The results cover training, post-training and agentic inference workloads.
Read the full article at NVIDIA
CoFabrix summarises and comments on this story. The original reporting belongs to NVIDIA.
Find out where your organization stands -- and what to do about it.