NVIDIA has topped MLPerf benchmark again — H100 Tensor Core GPUs running DGX H100 systems have aced every per test of AI inferencing in the latest round.
NVIDIA’s AI platform has consistently shown leadership across both training and inference since the inception of MLPerf, the gold standard for AI performance as an independent, third-party benchmark.
“Three years ago when we introduced A100, the AI world was dominated by computer vision. Generative AI has arrived. The industry is working hard on new advances in safe and trustworthy Generative AI. Hopper is enabling this essential work,” said Jensen Huang, Founder and CEO of NVIDIA.
Thanks to software optimisations, H100 delivered 54 percent more performance than six months ago.
Making its debut submission in MLPerf Inference v3.0, NVIDIA L4 Tensor Core GPU delivers over three times more performance than the prior-generation T4. L4 performs up to 10 times faster image decode, up to 3.2 times faster video processing and more than four times faster graphics and real-time rendering performance.
NVIDIA also made its debut submission in the network division, delivering up to 96 percent of the maximum local performance on BERT and 100 percent on ResNet-50.
Other results include improvements of up to 63 percent in energy efficiency and 81 percent in performance for the low-power Jetson AGX Orin module compared with its results from this time in 2022, while the Jetson Orin NX 16G delivered up to 3.2 times the performance of its Xavier NX predecessor.