Microsoft Azure customers can now flip to the most recent NVIDIA accelerated computing expertise to coach and deploy their generative AI purposes.

Accessible in the present day, the Microsoft Azure ND H100 v5 VMs utilizing NVIDIA H100 Tensor Core GPUs and NVIDIA Quantum-2 InfiniBand networking — allows scaling generative AI, excessive efficiency computing (HPC) and different purposes with a click on from a browser.

Accessible to clients throughout the U.S., the brand new occasion arrives as builders and researchers are utilizing massive language fashions (LLMs) and accelerated computing to uncover new client and enterprise use instances.

The NVIDIA H100 GPU delivers supercomputing-class efficiency by means of architectural improvements, together with fourth-generation Tensor Cores, a brand new Transformer Engine for accelerating LLMs and the most recent NVLink expertise that lets GPUs discuss to one another at 900GB/sec.

The inclusion of NVIDIA Quantum-2 CX7 InfiniBand with 3,200 Gbps cross-node bandwidth ensures seamless efficiency throughout the GPUs at huge scale, matching the capabilities of top-performing supercomputers globally.

Scaling With v5 VMs

ND H100 v5 VMs are perfect for coaching and operating inference for more and more advanced LLMs and laptop imaginative and prescient fashions. These neural networks drive essentially the most demanding and compute-intensive generative AI purposes, together with query answering, code era, audio, video and picture era, speech recognition and extra.

The ND H100 v5 VMs obtain as much as 2x speedup in LLMs just like the BLOOM 175B mannequin for inference versus earlier era cases, demonstrating their potential to additional optimize AI purposes.

NVIDIA and Azure

NVIDIA H100 Tensor Core GPUs on Azure present enterprises the efficiency, versatility and scale to supercharge their AI coaching and inference workloads. The mix streamlines the event and deployment of manufacturing AI with the NVIDIA AI Enterprise software program suite built-in with Azure Machine Studying for MLOps, and delivers record-setting AI efficiency in industry-standard MLPerf benchmarks.

As well as, by connecting the NVIDIA Omniverse platform to Azure, NVIDIA and Microsoft are offering a whole bunch of tens of millions of Microsoft enterprise customers with entry to highly effective industrial digitalization and AI supercomputing assets.

Study extra about new Azure v5 cases powered by NVIDIA H100 GPUs.