Spring over navigation

Azure NC A100 v4 VMs for AI now generally available

Udgivet den 1 juni, 2022

Senior Program Manager, Azure HPC and AI

AI is revolutionizing the world we live in—from the way we entertain ourselves, to the products and services that we consume, to the way we care for our bodies, and how we go about our daily work. Organizations are leveraging the power of AI to transform our lives by accelerating superior product innovations, increasing organization competitiveness no matter their size or available resources, and immersing us into more amazing, photo-realistic virtual worlds in movies and games.

At Microsoft, our mission is to empower every person and every organization on the planet to achieve more. With the power and scalability available through Microsoft Azure, we provide the compute tools and capabilities for all organizations no matter their size or resources to do more, faster. AI is a key tool to help organizations innovate and create new capabilities, discover new insights and deliver superior products and services. 

At Microsoft, our NC series virtual machines (VMs) allow our customers and partners access to almost limitless AI hardware infrastructure with Microsoft tools and services so that they can be productive quickly and more easily.

Here are some examples of what our customers are doing with the power of Azure AI:

“As scale complexity of advanced node integrated circuits increases, we are committed to provide our customers with innovative simulation technologies to reduce design and signoff cycles. Synopsys PrimeSim Continuum Solution delivers 10 times faster simulation with signoff accuracy using Azure’s NC A100 v4 virtual machines. The flexible form factor of NC A100 v4 VMs allow our customers to choose one, two, or four GPUs to fit the circuit size with the expected speedup of running on GPU.”—Hany Elhak, Sr. Director, Circuit Simulation Solutions, Synopsys

“In the era of large language models that often require tens to hundreds of Petaflops of GPU resources, how can a research lab in academia stay competitive? I think the answer is in cloud computing. Recently our strategy has been rigorously testing each new idea in a small scale until we have clear evidence that it works. And once we are sufficiently confident with it, we can radically scale it up with Azure’s high-end GPU virtual machines such as NC A100 v4 equipped with NVIDIA A100 80GB Tensor Core GPUs. This strategy has allowed us to become an active player in research areas that were formerly deemed to be only for industry labs.”—Minjoon Seo, Assistant Professor and Director of Language and Knowledge Lab, KAIST

For our MCity initiative, Touchcast leverages cloud-based rendering on Microsoft Azure GPU-optimized virtual machines. In our "Metaverse City", Touchcast delivers photorealistic 3D environments that bring organizations, employees, and customers together in a brand-new digital world. Best of all, MCity transports any Microsoft Teams user into stunning and immersive metaverse venues for meetings, presentations, workshop sessions, shopping experiences, and more."—Edo Segal, CEO and Founder, Touchcast

Introducing the new NC A100 v4 series virtual machine, now generally available

We are excited to announce that Azure NC A100 v4 series virtual machines are now generally available. These VMs, powered by NVIDIA A100 80GB PCIe Tensor Core GPUs and 3rd Gen AMD EPYC™ processors, improve the performance and cost-effectiveness of a variety of GPU performance-bound real-world AI training and inferencing workloads. Use the NC A100 v4 series for workloads like object detection, video processing, image classification, speech recognition, recommender, autonomous driving reinforcement learning, oil and gas reservoir simulation, finance document parsing, web inferencing, and more.

For AI training workloads, customers will experience between 1.5 to 3.5 times the performance boost compared to the previous NC generation (NCv3) with NVIDIA Volta architecture-based GPUs. Similar performance applies to AI Inference workloads.  Moreover, customers will experience between 1.5 to five times performance boost with seven independent GPU instances on a single A100 GPU through the multi-instance GPU (MIG) feature. Customers will experience increased performance gains with smaller batch sizes.

Increased GPU VM performance instances and more importantly, greater control over the GPU resources, means our customers can gain insights faster, innovate faster, and better utilize their available resources to do more with less. This means companies can diagnose cancer faster to save more lives, provide richer, more realistic virtual worlds for people to explore, and accelerate discoveries of cures for future potential communicable diseases.

The NC A100 v4-series offers three classes of virtual machines, ranging from one to four NVIDIA A100 80GB PCIe Tensor Core GPUs. It is more cost-effective than ever before, while still giving customers the options and flexibility they need for their workloads. We can’t wait to see what you’ll build, analyze, and discover with the new Azure NC A100 v4 platform.

Learn more