AMD and Azure

Azure Das v5 and Dav4 series VMs featuring AMD EPYC™ processors

Up to 96 vCPUs, 384 GBs of RAM, and 2,400 GBs of SSD-based storage

Example use cases: Enterprise grade apps, databases, web servers, dev/test

Azure Eas v5 and Eav4 series VMs featuring AMD EPYC™ processors

Up to 96 vCPUs, 672 GBs of RAM, and 2,400 GBs of SSD-based storage

Example use cases: Large in-memory business critical workloads like analytics, business critical apps, and relational databases such as SQL Server

Azure Lasv3 series VMs featuring AMD 3rd Generation EPYC™ 7763v processor

Up to 19.2TB NVMe storage and Premium SSD support for high throughput, low latency, and large local disk storage for database apps

Example use cases: NoSQL databases, big data analytics, large transactional databases, and data warehousing

Get affordable, high-performing modern desktop experiences from the cloud with Azure NVv4 instance VMs—ideal for financial analysis, architecture, engineering, and design applications. Traditional CPU-only VMs often struggle to keep up with these apps, frustrating users and reducing productivity, but until now, deploying GPU-accelerated, on-premises virtual environments has been too costly.

Powered by AMD Radeon Instinct™ GPUs and the industry’s first SR-IOV-based fractional GPU virtualization architecture, NVv4 instances offer a choice of four configurable options to deliver the right amount of GPU capability at the right price for any enterprise workload.

NVv4 instances provide a zero-cost license GPU partitioning on discrete GPUs in Azure.

Learn more about GPU partitioning on Azure NVv4 instances:

NVv4 datasheet

GPU virtualized architecture with the MxGPU whitepaper

NVv4 pricing

Ideal for designers, architects, engineers, and other professionals who use resource- and performance-intensive applications, NVv4-based Azure virtual desktops bring the full power of traditional workstations into the cloud, regardless of the end-point device used or if working onsite or remotely.

Designed with major pro-CAD architecture, engineering, and construction (AEC) and M&E applications, Azure NVv4 instances are ISV-certified to ensure optimal performance and reliability. Zero-cost-license GPU partitioning provides a discrete GPU experience even when fractional GPU options are deployed, making it possible to balance performance, mobility, and cost.

For more information on the benefits of NVv4 for industries that requires workstation-class experiences, view the latest solution briefs:

NVv4 in design and manufacturing

NVv4 in AEC

NVv4 in education

NVv4 in finance

To accelerate scientific and commercial research, discovery, and innovation, Azure and AMD have collaborated to offer extensive on-demand supercomputing resources with highly scalable HPC-optimized VMs.

Azure HBv3 VMs are designed to deliver leadership-class performance, message-passing interface (MPI) scalability, and cost efficiency for a variety of real-world HPC workloads such as CFD, finite element analysis, computational geoscience, rendering, weather simulation, financial risk analysis, and RTL simulation.

Learn more about Azure HBv3:

Azure HBv3 GA announcement

Performance and scalability results

For a wide range of HPC applications including computational fluid dynamics and weather simulation, Azure HBv2 VMs feature 2nd generation AMD EPYC™ processors, providing 45-50 percent more memory bandwidth than comparable x86 alternatives. HBv2 VMs also feature the cloud’s first 200 Gbps HDR InfiniBand from Mellanox for ultra-low latency and accelerated message transfer.

Learn more about Azure HBv2:

Benchmarking results: Weather research and forecasting (WRF) in Azure

GitHub: Get up and running with an Azure HPC cluster

Azure HPC Cache: Run flexible, file-based workloads in Azure

Blog: HBv2 eclipses 80,000 core milestone

ServetheHome: Microsoft Azure HBv2 HPC Features AMD EPYC™ 7002 and Mellanox Report

Azure DCasv5 and ECasv5 series VMs allow you to move your existing workloads to Azure and make them confidential without changing any code. With the latest advancement in AMD SEV technologies called Secure Nested Paging (SEV-SNP), the contents of entire VMs are opaque to cloud administrators for secure and isolated computation. The runtime state of these VMs is fully encrypted, protecting your data even when it's in use. The keys used for this RAM encryption are generated inside the CPU and never leave it.

Learn more about the Azure DCasv5 and ECasv5 series:

Azure DCasV5 and ECasv5 preview announcement

Modernize aging server and storage infrastructure, consolidate virtualized workloads, and gain cloud efficiencies on-premises. AMD and Azure hyperconverged solutions bring together software-defined compute, storage, and networking on industry-standard x86 servers and components. Validated hardware options make it possible for enterprises to gain the benefits of Azure Stack HCI solutions, powered by the performance of AMD EPYC™ processors, from within their own, purpose-built datacenters.