Nvidia introduced yesterday that its upcoming H100 “Hopper” Tensor Core GPU set new efficiency information throughout its debut within the industry-standard MLPerf benchmarks, delivering outcomes as much as 4.5 occasions quicker than the A100, which is at the moment Nvidia’s quickest manufacturing AI chip.
The MPerf benchmarks (technically referred to as “MLPerfTM Inference 2.1”) measure “inference” workloads, which reveal how properly a chip can apply a beforehand skilled machine studying mannequin to new information. A gaggle of {industry} corporations often called the MLCommons developed the MLPerf benchmarks in 2018 to ship a standardized metric for conveying machine studying efficiency to potential prospects.
Specifically, the H100 did properly within the BERT-Giant benchmark, which measures pure language-processing efficiency utilizing the BERT mannequin developed by Google. Nvidia credit this explicit consequence to the Hopper structure’s Transformer Engine, which particularly accelerates coaching transformer fashions. Because of this the H100 might speed up future pure language fashions just like OpenAI’s GPT-3, which may compose written works in many various kinds and maintain conversational chats.
Learn 2 remaining paragraphs | Feedback