Microsoft Touts Windows Server 2025's GPU Partitioning Feature

Microsoft won its place at the near-top of the generative AI food chain by inking early partnerships with companies like ChatGPT steward OpenAI and Nvidia, whose powerful GPUs are able to support resource-intensive AI model training, deep learning inferencing and other AI tasks.

Windows Server 2025, recently released to public preview, promises to be a major beneficiary of these relationships -- specifically, the one with Nvidia.

In Microsoft's late-May announcement of the preview release, for instance, it described the improvements coming to Windows Server 2025 as spanning "AI, performance and scale."

"Azure hosts some of the world's largest workloads that push the limits of CPU and memory capabilities to process huge data sets across distributed environments," the company said at the time. "With the growth of AI and machine learning, GPUs have become a key part of cloud solutions because they're great at performing many parallel operations on large data. Windows Server 2025 brings you many of these advantages across GPUs, storage, networking, and scalability."

In an article last week titled "Introducing GPU Innovations with Windows Server 2025," Microsoft went into more detail about these GPU benefits.

"GPUs are essential for AI due to their parallel processing capabilities and highly scalable architecture," said the article authored by Afia Boakye and Rebecca Wambua on Microsoft's Tech Community site, discussing Windows Server 2025 Datacenter and Azure Stack HCI 24H2.

"Using the upcoming OS releases, Microsoft's customers can provide an entire GPU to a VM, which can run either Linux or Server, in a failover cluster using discrete device assignment (DDA)," they said. "This means that mission-critical AI workloads can easily run in a clustered VM and, upon an unexpected fault or a planned move, the VM will restart on another node in the cluster, using a GPU on that node.

Microsoft is touting "GPU partitioning" as a key differentiator coming to Windows Server 2025. GPU partitioning helps users partition a supported GPU and assign those partitions to different VMs in a failover cluster, resulting in multiple VMs that can share a single physical GPU, giving each VM an isolated fraction of the physical GPU's capabilities.

"Further, due to a planned or unplanned move, the VMs will restart on different nodes in the cluster, using GPU partitions on those different nodes," Microsoft said. "Besides enabling clustered VMs to use GPU-P, the upcoming OS releases are bringing live migration to VMs using GPU-P. Live migration for GPU-P enables customers to balance mission-critical workloads across their fleet and to conduct hardware maintenance and software upgrades without stopping their VMs."

More information about Windows Server 2025 is available here.

About the Author

David Ramel is an editor and writer for Converge360.


comments powered by Disqus

Subscribe on YouTube