Yesterday on August 29th, Microsoft announced that a variety of Azure virtual machines are now enabled with NVIDIA GPUs. This means that users can now access GPU-optimized software for use with Azure instances and high-performance computing workloads.
Microsoft is specifically mentioning that the support for the NVIDIA GPUs in the cloud could be good for those using deep learning software. This includes tested and approved containers for the Microsoft Cognitive Toolkit, TensorFlow, PyTorch, and NVIDIA TensorRT. NVIDIA has also created optimal software stack for each of these for optimal performance with their GPUs, including required operating system patches, NVIDIA deep learning libraries, and the NVIDIA CUDA Toolkit.
A new NVIDIA GPU Cloud Image for Deep Learning image is now available in the Azure Marketplace for those who are interested. You can also sign up now for a free account and then pull the containers into your Azure instance. For more information, feel free to sign up for the webinar on October 2nd.
Microsoft also announced General Availability of Azure CycleCloud, "a tool for creating, managing, operating, and optimizing HPC clusters of any scale in Azure." According to the blog post, "Azure CycleCloud’s role-based policies and governance features make it easy for their organizations to deliver the hybrid compute power where needed while avoiding runaway costs."