Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
The latest MLPerf Training v4.0 benchmark results from MLCommons have been released, highlighting significant advancements in AI training performance. NVIDIA's accelerated computing platform, powered by Hopper GPUs and Quantum-2 InfiniBand networking, delivered exceptional results, more than tripling large language model (LLM) performance compared to last year. The NVIDIA H100 was pivotal in these achievements, with 11,616 units utilized. Intel's Gaudi 2 AI Accelerator also performed notably, being the only alternative benchmarked against NVIDIA's H100. The results showcase up to 80% gains in AI performance, emphasizing the importance of both hardware and software innovations.
Can someone explain what @realGeorgeHotz did here? MLPerf is a benchmark suite that is used to evaluate training and inference performance of on-premises and cloud platforms, usually used for Nvidia GPUs? So he got an AMD GPU to benchmark on an Nvidia benchmark meaning he… https://t.co/KnUanLWapo
Exciting news for the #AI community! Learn how our @MLPerf results demonstrate the choice that Gaudi 2 AI Accelerators bring to enterprises and customers. #GenAI #Intel https://t.co/yFZFHj81ou
👀 In #MLPerf Training v4.0, we set new #generativeAI training performance records and continued to deliver the highest performance on every workload✨🏆 Technical Deep Dive ➡️https://t.co/JokNZtD7D8 Performance delivered using the full stack of NVIDIA software and hardware. https://t.co/mOQXRAOzbM
