Azure Extends Access to NVIDIA GPU Cloud for Deep Learning

Graphics Processing Units or GPUs offer outstanding performance for AI (Artificial Intelligence) and HPC (High Performance Computing) workloads. Microsoft’s Azure platform is already supporting high performance computing, deep learning and advanced scale workloads by providing a variety of virtual machines enabled with NVIDIA GPUs. Starting this week, Azure users and cloud developers have access to a new way of accelerating their AI and HPC workflows with powerful GPU-optimized software. Microsoft and NVIDIA’s new collaboration will make it easier for developers to access high-performance GPUs in the cloud.

Both Microsoft and NVIDIA are targeting data scientists, developers and researchers who will be given extensive access to GPU compute resources. Microsoft has added a new level of support to its Azure platform for facilitating NVIDIA GPU projects. This will benefit those who are running deep learning and high performance computing workloads. Users have a choice of 35 different pre-configured containers with GPU acceleration for deep learning software, HPC applications, HPC visualization tools and more. NVIDIA’s containers work seamlessly across different Azure instance types, even if the number of GPUs across containers are not uniform.

Launched in the beginning of this year, NVIDIA’s GPU Cloud is driven by Volta and the Tensor Core GPU architecture, and its container registry supports deep learning tools like TensorFlow, PyTorch, and Microsoft’s Cognitive Toolkit. With Microsoft’s commitment to make Azure the cloud choice for HPC, this new partnership is expected to bring significant changes in software deployment in the next few years. Additionally, Microsoft has announced the general availability of Azure CycleCloud, which is defined as a tool for creating, managing, operating, and optimizing HPC clusters of any scale in Azure.