Azure VMs with AMD MI300 GPU Accelerators at Microsoft Conference

Nov. 15, 2023 — Today at Microsoft Ignite, AMD (NASDAQ: AMD) and Microsoft featured AMD products, including the upcoming Instinct MI300X GPU accelerator, EPYC and Ryzen CPUs with AI engines.

Microsoft’s new Azure ND MI300x v5 Virtual Machine (VM) series optimized for AI workloads features the MI300X, making Azure the first cloud to use the new accelerator. The new VMs are part of Microsoft’s  infrastructure supporting AI for enterprises everywhere giving customers more choice in efficiency and scalability.

In a blog on the VM, Microsoft said it has 1.5 TB of high bandwidth memory (HBM) that leverages the MI300X. “What differentiates a ND MI300X v5 VM from our existing family of ND-series VMs is the inclusion of 8 x AMD Instinct MI300X GPUs interconnected via Infinity Fabric 3.0 in each VM,” Microsoft said. “MI300X offers industry leading HBM capacity and bandwidth, 192 GB of HBM3 memory per GPU capable of speeds up to 5.2 TB/s.”

The VMs include:

  • 400 Gb/s NVIDIA Quantum-2 CX7 InfiniBand per GPU with 3.2Tb/s per VM
  • 4th Gen Intel Xeon Scalable processors
  • PCIE Gen5 host-to-GPU interconnect with 64GB/s bandwidth per GPU
  • 16 Channels of DDR5 DIMMs

HBM is essential for AI applications due to its high bandwidth, low power consumption, and compact size, Microsoft said. “It is ideal for AI applications that need to quickly process vast amounts of data. ND MI300X v5 has the most HBM capacity available, 1.5 TB per VM or 192 GB of HBM per GPU, the highest HBM capacity available in the cloud. This lets customers run the largest, most advanced AI models faster, and with fewer GPUs.”

4th Gen AMD EPYC processors are also now being used to run a new generation of general purpose, memory-intensive and compute-optimized VMs. These new VMs from Microsoft continue to showcase the growth and demand of AMD EPYC processors in the cloud and will provide better price options, up to 20% better performance for general purpose and memory-intensive VMs, and up to 2x the CPU performance for compute-optimized VMs versus the previous generation of AMD EPYC-powered VMs. The new VM series will be made available in public preview in the first quarter of 2024.

At Ignite, AMD is also highlighting the Azure NGads V620 series of VMs, now in general availability. It is a new addition to the Azure N-series family of GPU-accelerated VMs for visualization. These VMs are powered by multiple AMD technologies – AMD Radeon™ PRO V620 GPUs and 3rd Gen AMD EPYC CPUs. The NGads V620 series is intended for workloads that require relatively more GPU resources per VM, to support more demanding workloads or more users per VM. The V620 delivers a great experience for multiple classes of GPU workloads, including gaming, VDI and rendering. NG series users get an optimized driver through the new AMD Software: Cloud Edition.

“AMD is fostering AI everywhere – from the cloud, to the enterprise and end point devices – all powered by our CPUs, GPUs, accelerators and AI engines,” said Vamsi Boppana, Senior Vice President, AI, AMD. “Together with Microsoft and a rapidly growing ecosystem of software and hardware partners, AMD is accelerating innovation to bring the benefits of AI to a broad portfolio of compute engines, with expanding software capabilities.”

AMD also showcased the ongoing momentum of Ryzen AI, the first dedicated AI accelerator available on an x86 processor. With over 50 systems available now with Ryzen 7000 Series processors with Ryzen AI built in, there are millions of AMD AI PCs available in market. AMD and Microsoft have already jointly enabled Windows Studio Effects on Ryzen AI PCs, and AMD continues to work with software developers, to bring expanding AI workloads and generative AI experiences to consumers.