Run Siemens NX using AMD Radeon PRO V710 GPUs on Azure Virtual Desktop

Overview

This article presents a proof of concept (PoC) that validates Siemens NX 2506 on Azure Virtual Desktop (AVD) using AMD Radeon PRO V710 GPUs on Azure NVads V710 v5 infrastructure. The assessment evaluates how Siemens NX performs in shared virtual workstation environments with multiple concurrent users.

Testing used Standard_NV12ads_V710_v5 and Standard_NV24ads_V710_v5 virtual machines to measure application responsiveness, graphics performance, scalability, and overall user experience. The validation included Siemens NX Automated Test Suite (ATS) certification testing, single-host and multihost deployment scenarios, and concurrent user scalability assessments.

The findings help organizations understand the suitability of AVD for engineering and computer-aided design (CAD) workloads powered by AMD Radeon PRO V710 GPUs and provide guidance for selecting virtual machine configurations based on workload requirements.

Current challenges

Organizations that provide high-performance Siemens NX engineering environments often face the following infrastructure cost, security, and scalability challenges:

• Costly, high-maintenance workstation hardware: Traditional deployments require expensive workstations with dedicated graphics capabilities, along with ongoing maintenance and frequent upgrades to support complex 3D modeling, simulation, and rendering workloads.

• Secure access for remote and distributed teams: Supporting remote and distributed engineers adds complexity, because organizations must provide secure access to applications and design data without compromising performance or intellectual property.

• Inconsistent experiences and limited scalability: Managing multiple physical workstations can lead to inconsistent user experiences, increased administrative effort, and difficulty scaling resources to meet changing project requirements.

Key benefits

By using AVD and AMD Radeon PRO V710 GPU-enabled virtual machines, engineering teams can access centrally managed Siemens NX desktops and applications.

• Validated Siemens NX workloads: Siemens NX 2506 completed ATS validation testing on AVD environments powered by AMD Radeon PRO V710 GPUs.

• GPU-accelerated engineering workloads: Azure NVads V710 v5 virtual machines provide graphics acceleration for Siemens NX visualization, modeling, display operations, and engineering workflows.

• Remote access with centralized data: Engineers can securely access Siemens NX workloads through AVD while project data remains centralized in Azure-hosted infrastructure.

• Scalable deployment options: The PoC tested configurations ranging from a single 12-vCPU session host to two 24-vCPU session hosts. The two-host, 24-vCPU configuration completed testing with 30 concurrent users.

• Centralized management and administration: AVD simplifies management of Siemens NX applications, licensing infrastructure, user profiles, and security policies.

Prerequisites

Before deploying and validating Siemens NX 2506 on AVD, ensure the following requirements are in place:

1. Azure environment

   • An active Azure subscription.

   • An AVD deployment including a host pool, application group, and workspace.

   • A Standard_NV12ads_V710_v5 virtual machine configured as an AVD session host, with 64 GiB of memory, one-half GPU, and 12 GB of GPU memory.

   • A Standard_NV24ads_V710_v5 virtual machine configured as an AVD session host, with 128 GiB of memory, one GPU, and 24 GB of GPU memory.

2. Operating system and GPU drivers

   • Windows 11 Enterprise multi-session operating system installed on both the AVD session host and Siemens License Server virtual machines.

   • AMD Radeon PRO V710 GPU driver version 32.0.21017.2052 installed and configured for hardware-accelerated graphics.

3. Siemens NX software

   • Access to the Siemens NX 2506 installation media downloaded from the Siemens Software Support Portal.

4. Licensing infrastructure

   • A valid Siemens NX license file.

   • A dedicated Windows virtual machine configured with Siemens License Server Manager for centralized license administration.

5. User access and connectivity

   • Microsoft Entra ID user accounts assigned to AVD.

   • Microsoft Remote Desktop client installed on user devices.

   • Network connectivity between the AVD session host and the Siemens License Server.

6. GPU validation

   • Verify that the AMD GPU is correctly detected in Device Manager under Display adapters.

   • Confirm that Siemens NX 2506 launches successfully with GPU acceleration enabled.

Architecture

This architecture delivers Siemens NX as GPU-accelerated desktops and applications on AVD. Users connect from any device through the Remote Desktop client, and Microsoft Entra ID authenticates each session. Azure Virtual Network routes traffic to pooled, multisession session hosts that run Siemens NX on AMD Radeon PRO V710 GPUs, while centralized licensing and storage services provide Siemens NX licenses, user profiles, and shared engineering data.

Workflow

The following steps outline the end-to-end process for accessing and running Siemens NX workloads on AVD:

1. User access to AVD: Users connect to AVD by using the Remote Desktop client from their local devices to access published Siemens NX applications or a dedicated desktop session.

2. User authentication and access: Microsoft Entra ID authenticates users and provides secure access to assigned AVD resources and Siemens NX workloads.

3. Session routing through Azure networking: Azure networking and connectivity services securely route user sessions from the on-premises environment to AVD session hosts through the virtual network infrastructure.

4. GPU-enabled Siemens NX processing: GPU-enabled AVD session hosts run Siemens NX on Azure VMs powered by AMD Radeon PRO V710 GPUs, delivering high-performance graphics, visualization, and engineering application experiences for concurrent users.

5. Centralized management, licensing, and storage: Centralized management, licensing, and storage services provide secure access to Siemens NX licenses, user profiles, and engineering data while enabling simplified administration, scalability, and a consistent user experience across the environment.

Components

This reference architecture uses the following Azure components.

Azure Virtual Desktop is a desktop and app virtualization service that runs on Azure. Deliver a full Windows experience with Windows 11, Windows 10, or Windows Server. Use single-session host pools to assign a device to a single user, or use multisession host pools for scalability. Use it to provide users with a virtualized desktop environment for CAD workstations and let them access NX from anywhere. For more information, see Azure Virtual Desktop for the enterprise.

Azure Virtual Machines is an infrastructure as a service (IaaS) offering that provides on-demand, scalable compute without the need for physical hardware maintenance. Use virtual machines to host the NX applications.

Azure managed disks provide persistent and reliable block storage for the Azure virtual machines that run Siemens NX workloads. Use managed disks to store the operating system, Siemens NX application files, project data, and user files while delivering consistent performance and simplified storage management through Azure-managed infrastructure.

Azure Virtual Network is a service that facilitates secure communication between Azure resources, the internet, and on-premises networks. In this architecture, use it to create a secure network infrastructure for reliable communication.

Microsoft Entra ID provides on-premises directory synchronization and single sign-on features. Use Microsoft Entra ID to manage and authenticate users and provide access to Siemens NX on AVD from the Remote Desktop client.

Remote Desktop client lets you connect to Windows desktops and apps over a network by using the Remote Desktop Protocol (RDP). Connect to Windows hosted in the cloud from AVD, to Remote Desktop Services on-premises, and to remote PCs.

Terminology

This section defines the key AVD concepts used in this architecture.

Session hosts are GPU-enabled Azure virtual machines. Host pools are a collection of Azure virtual machines that are registered to AVD as session hosts. For a consistent user experience, all session host virtual machines in a host pool should come from the same image.

A session host configuration is a subresource that specifies the settings for session hosts in a host pool. The configuration persists for the host pool's lifecycle and aligns with its session hosts.

An application group controls access to a full desktop or a logical grouping of applications that are available on session hosts in a single host pool. AVD supports multiple application groups that publish full session desktops or collections of applications, such as Siemens NX, Microsoft 365 apps, and other CAD applications. You can assign users to multiple application groups across multiple host pools, which lets you vary the applications and desktops users can access. When you create an application group, choose one of these types:

• Desktop: Users access the full Windows desktop from a session host.

• RemoteApp: Users access individual applications you select and publish to the application group. Available with pooled host pools only.

A workspace is a logical grouping of application groups. Each application group must be associated with a workspace for users to see the desktops and applications published to them. The Remote Desktop application displays them. You can only assign an application group to a single workspace.

Use network security groups to limit access to subnets within the Azure network. For deployments of Siemens NX on AVD, use network security groups to secure the network infrastructure so that only authorized traffic can access the session desktop.

A cloud administrator creates a VM image that includes the operating system, Siemens NX 2506, AMD Radeon PRO V710 graphics drivers, Siemens NX licensing configuration, user profile data, and other required files. Use the VM image to create session hosts and provide a consistent AVD deployment for Siemens NX workloads. Create an image from a VM and then use it to create multiple VMs. For images stored in an Azure Compute Gallery (formerly Shared Image Gallery), use VMs that already have accounts (specialized), or generalize the VM before creating the image to remove machine accounts and other machine-specific information. To generalize a VM, see Generalize a VM.

Remote Desktop client and connection

1. Download and install the Microsoft Remote Desktop client on your endpoint device to access AVD resources.

2. Sign in with your organizational credentials to view the published AVD workspace, including the Siemens NX application and assigned desktop resources.

   

3. Launch a virtual desktop session or the Siemens NX RemoteApp from the workspace to open a GPU-accelerated engineering environment hosted in Azure.

4. Create, modify, and save Siemens NX designs directly in centralized Azure-hosted storage for secure data management and team collaboration.

Map a shared drive for Siemens NX on Azure Virtual Desktop

For this validation, Siemens NX project data was stored on an Azure managed disk attached to the AVD session host. Using a dedicated managed disk provided centralized storage for model files, assemblies, drawings, and other engineering data, enabling users to access project files from a common location instead of storing them on individual profiles or local workstation storage. The managed disk was available in every user session as a shared drive for all concurrent users on the host.

Storing Siemens NX data on a managed disk helps maintain a consistent repository for engineering projects, simplifies data management, and ensures that project files remain available within the Azure environment. Designers can save, open, and modify files directly from the shared storage location, reducing file duplication and improving collaboration across engineering teams.

Mount the managed disk

1. Attach the Azure managed disk to the AVD virtual machine.
2. Sign in to the session host by using an administrative account.
3. Open Disk Management.
4. Initialize and format the newly attached disk if required.
5. Assign a drive letter (for example, D:).
6. Create a folder structure for Siemens NX project data.
7. Configure NTFS permissions to allow authorized users or engineering groups to access the shared location.
8. In Siemens NX, save and access projects from the mounted drive.

Alternative storage options

While this study uses a managed disk, organizations with larger engineering teams or advanced collaboration requirements might consider other Azure storage services, including:

• Azure Files for Server Message Block (SMB)-based shared file access.

• Azure NetApp Files for high-performance CAD and engineering workloads.

• Windows File Server on Azure for traditional file-sharing requirements.

Choose the appropriate storage solution based on factors such as performance requirements, scalability, user concurrency, data management practices, and business continuity objectives.

Validate Siemens NX on Azure Virtual Desktop

Get the NX applications, NX Automated Test Suite (NXATS), and NX Certification Pack (NXCP) from Siemens Support. To validate Siemens NX on AVD, use NXATS in manual or automated mode and NXCP in automated or interactive mode.

Get test suites and installation instructions from the Global Technical Access Center (GTAC).

Validation results

ATS validates graphics rendering behavior and runs a broad collection of graphics-intensive operations. Although ATS doesn't cover every NX feature, it provides a comprehensive assessment of rendering performance and graphics subsystem functionality.

For this proof of concept, success meant that all Siemens NX sessions completed the ATS operations without the application hanging or stalling. The 24-vCPU two-host configuration was tested with up to 30 concurrent users, as shown in the following results tables.

The following section lists the ATS graphics operations tested and their execution times across the tested configurations. The sequence matches the order shown in the results charts.

12-vCPU AMD V710

The following tables compare ATS operation runtimes for the 12-vCPU AMD V710 configuration across increasing numbers of concurrent users in single-host and two-host environments.

12-vCPU single-host results

12-vCPU two-host results

12-vCPU results analysis

The 12-vCPU AMD V710 single-host configuration maintained stable runtimes across most operations at lower user loads. Visualization remained the most resource-intensive operation, with longer runtimes than the other ATS operations.

The 12-vCPU two-host configuration used breadth-first load balancing, which distributes new sessions across available hosts. Siemens NX sessions maintained the expected level of concurrency for up to 12 concurrent users. When the workload increased to 14 concurrent users, NX successfully launched all concurrent sessions; however, concurrency wasn't consistently maintained across all sessions. Some sessions completed later than others while waiting for available compute resources, such as CPU scheduling and system resource allocation.

Across matched user loads, the two-host configuration recorded lower runtimes for most ATS operations, although not every operation improved. At the highest tested load, all sessions completed successfully, and no application hangs, failures, or functional issues were observed from the end-user perspective. The observed delays were related to compute resource availability under increased load rather than any application-level limitation.

24-vCPU AMD V710

The following tables compare ATS operation runtimes for the 24-vCPU AMD V710 configuration across increasing numbers of concurrent users in single-host and two-host environments.

24-vCPU single-host results

24-vCPU two-host results

24-vCPU results analysis

The 24-vCPU AMD V710 single-host configuration delivered strong performance at lower and moderate user loads. As concurrency increased, runtimes rose noticeably, with significant spikes in resource-intensive operations such as visualization, indicating higher resource contention.

In contrast, the multihost configuration distributed the workload more efficiently, resulting in more stable and scalable performance at higher user counts and better handling of resource-intensive operations. Although both configurations performed similarly under lighter workloads, the multihost configuration provided better scalability and reliability for higher concurrency and more demanding scenarios. However, scaling out didn't consistently reduce runtimes at lower user concurrency.

Validation results summary

Selected 12-vCPU and 24-vCPU results

The following chart compares four selected test points: a 12-vCPU single-host configuration with 6 users, a 12-vCPU two-host configuration with 12 users, and 24-vCPU single-host and two-host configurations with 12 users.

The comparison shows that the 24-vCPU configuration performs faster and handles higher workloads better than the 12-vCPU setup. While both perform similarly for smaller tasks, the 24-vCPU consistently delivers lower runtimes and more stable results as user load and data size increase. Heavy operations like visualization show significant improvement with 24 vCPU, with fewer slowdowns and spikes. Overall, 24 vCPU is more efficient and scalable, making it a better choice for high-demand environments.

For the tested Siemens NX workloads, increasing compute capacity from 12 vCPUs to 24 vCPUs and distributing user sessions across multiple Azure Virtual Desktop session hosts improved overall application responsiveness and reduced workload execution times. The most significant gains were observed in visualization, display manipulation, and model analysis operations. These results indicate that organizations supporting concurrent Siemens NX users can improve performance and scalability by combining larger VM sizes with multi-session-host deployment strategies.

Scenario details

The following scenario shows how a manufacturing organization can apply the architecture and validation findings. The organization deploys Siemens NX on AVD to provide engineers with secure access to CAD, product engineering, and product lifecycle management (PLM) workflows from any location. Previously, its engineering teams relied on locally installed workstations that required significant hardware investments, complex lifecycle management, and time-consuming upgrades to support growing design requirements.

In this scenario, the organization centralizes its engineering environment while giving users access to high-performance virtual workstations through Azure. Engineers can create, modify, analyze, and visualize complex product designs by using Azure GPU-enabled virtual machines across office, remote, and hybrid work environments. Azure Files, Azure NetApp Files, or managed disks provide centralized access to design data, while Microsoft Entra ID and Azure networking services control access to engineering resources. This approach simplifies workstation management, supports collaboration across distributed teams, and enables the organization to scale engineering resources as project requirements evolve.

Potential use cases

Siemens NX on AVD supports product design, engineering, and manufacturing workflows that require secure access to high-performance CAD environments. Consider this architecture for the following use cases:

• Product design and mechanical engineering for creating and managing complex 3D parts, assemblies, and technical drawings.

• Large assembly visualization and design reviews for enabling engineering teams to work with complex product models and collaborate across locations.

• Computer-aided manufacturing (CAM) workflows for numerical control (NC) programming, toolpath generation, and manufacturing process planning.

• PLM integration for accessing design data, managing revisions, and supporting engineering change processes.

• Remote and distributed engineering teams that require secure access to Siemens NX without relying on high-end local workstations.

• Digital prototyping and virtual product development for helping organizations reduce design iterations and accelerate product development cycles.

• Graphics-intensive engineering workloads that benefit from Azure GPU-enabled virtual machines for model visualization and interaction.

This architecture is applicable across a wide range of industries, including automotive, aerospace and defense, industrial equipment, energy and utilities, shipbuilding, high-tech electronics, consumer products, medical devices, and engineering services organizations. By hosting Siemens NX on AVD, organizations can deliver a centralized, secure, and scalable engineering environment while simplifying workstation management and supporting modern hybrid work requirements.

Considerations

The following considerations align with the pillars of the Azure Well-Architected Framework, which is a set of guiding tenets used to improve the quality of a workload. For more information, see Well-Architected Framework.

Reliability

Reliability helps ensure that your application can meet the commitments that you make to your customers. For more information, see Design review checklist for Reliability.

AVD provides Microsoft-managed brokering, gateway, web, and resource-directory services. You're responsible for the resilience of session hosts, virtual-network connectivity, identities, profile storage, Siemens NX application images, engineering data, and licensing. For production, deploy at least two identically configured GPU session hosts and distribute them across availability zones where the selected NVads V710 v5-series capacity is available. Use separate failure domains or availability sets when zones aren't available. Enable drain mode before maintenance, keep spare capacity for host or zone failure, and validate that disconnected users can reconnect to a healthy host without losing access to their profiles or NX project data.

Store FSLogix profile containers on highly available Azure Files or Azure NetApp Files storage so users can retain personalization when sessions move between hosts. Protect engineering models, shared project data, golden images, and the Siemens License Server configuration with workload-appropriate backup and restore procedures. Avoid treating ephemeral session-host disks as the system of record.

For business-critical deployments, define recovery time and recovery point objectives, maintain a tested secondary-region design, and consider paired host pools with replicated profile and project storage. Azure Site Recovery can protect suitable VM-based dependencies, while FSLogix Cloud Cache can support profile availability across storage locations or regions. Regularly test host replacement, loss of a session host, profile-storage interruption, license-server recovery, and regional failover. Document the reconnection and failback process.

Security

Security protections guard against deliberate attacks and the misuse of your valuable data and systems. For more information, see Design review checklist for Security.

Apply defense in depth across identity, session hosts, networking, data, and operations. Assign Siemens NX desktops or RemoteApps through Microsoft Entra ID security groups, require multifactor authentication with Conditional Access, and evaluate device compliance, location, sign-in risk, and session frequency for the engineering user population. Separate end-user access from administration, grant Azure and AVD roles using least privilege, and use privileged access workflows for elevated tasks. AVD uses reverse connect, so session hosts don't require inbound RDP ports from the internet. Block direct public RDP and restrict administrative paths to approved management networks and tools.

Place session hosts, profile storage, engineering data services, and the Siemens License Server on dedicated subnets. Use network security groups and Azure Firewall to allow only required east-west and outbound traffic, including the approved Siemens licensing ports. Siemens NX is configured to use port 28000 as the default license server communication port. You can customize this port based on deployment, security, or networking requirements. Use private endpoints for profile and project storage and consider Azure Private Link for AVD when private connectivity is required. Harden and patch the Windows image, AMD driver, Siemens NX, and supporting agents through a controlled image lifecycle. Enable Microsoft Defender for Cloud and endpoint protection.

Centralize Windows, Entra ID, AVD, firewall, and license-server logs, and alert on unusual sign-ins, privilege changes, malware, and anomalous data access. Protect intellectual property by limiting clipboard, drive, printer, and USB redirection according to business need. For sensitive designs, evaluate screen-capture protection, watermarking, and data-loss-prevention controls. Encrypt data in transit and at rest. Manage secrets and automation credentials in Azure Key Vault. Regularly review user, administrator, service-principal, and license-server access.

Cost optimization

Cost Optimization focuses on ways to reduce unnecessary expenses and improve operational efficiencies. For more information, see Design review checklist for Cost Optimization.

If your workload requires more memory and fewer CPUs, use a constrained vCPU virtual machine size to reduce software licensing costs charged per vCPU.

The Azure calculator helps you estimate and optimize costs. For an estimated cost of the baseline architecture, see estimated cost. Your estimates might differ based on your implementation of Siemens NX on AVD.

Use the SKU recommendations section for guidance on selecting the right virtual machine SKUs. Contact the Siemens support team for the latest NX on Azure certification matrix and detailed SKU recommendations.

Operational excellence

Operational Excellence covers the operations processes that deploy an application and keep it running in production. For more information, see Design review checklist for Operational Excellence.

Standardize session host provisioning by maintaining a golden Siemens NX image in Azure Compute Gallery that includes the validated NX release, AMD GPU driver, required environment variables, and monitoring agents. Update it through a controlled image lifecycle. Automate host pool deployment, scaling plans, and configuration with infrastructure as code so environments remain consistent and repeatable.

Monitor the end-to-end experience with Azure Monitor and Log Analytics. Track session host health, GPU and CPU utilization, connection latency, FSLogix profile load times, and Siemens License Server availability. Establish alerts and operational runbooks for host replacement, driver updates, license-server recovery, and NX version upgrades. Validate changes in a nonproduction host pool before rolling them into production.

Performance efficiency

Performance Efficiency is the ability of your workload to scale to meet the demands placed on it by users in an efficient manner. For more information, see Design review checklist for Performance Efficiency.

The PoC results indicate that AVD running on AMD Radeon PRO V710 GPUs can support Siemens NX workloads across the tested configurations. Testing validated 30 concurrent Siemens NX users on two 24-vCPU GPU-enabled session hosts. The ATS operations completed without the application hanging or stalling, which met the success criterion for this PoC.

As user density increased, the multihost architecture sustained performance under higher concurrent workloads. At 20 concurrent users, distributing sessions across two 24-vCPU hosts reduced FaceAnalysis execution time by about 58% and VisualizationPerformance execution time by about 50% compared with the single-host configuration. For the full per-operation results, see Validation results. These results indicate that AVD is a flexible, high-performance platform for scaling Siemens NX deployments while optimizing infrastructure utilization and user productivity.

Scalability

Use pooled, multisession host pools to scale Siemens NX from a single GPU-enabled session host to multiple hosts as concurrent demand grows. The validation tested up to 30 concurrent users and showed that distributing 20 users across two hosts improved most graphics-intensive operations compared with placing all 20 users on one host. For the per-operation improvements, see Validation results summary. These results support scaling out before GPU, CPU, memory, or session limits create sustained contention.

Configure host-pool load balancing to match workload patterns. Breadth-first allocation spreads sessions across available hosts to preserve per-user graphics headroom, while depth-first allocation fills hosts more densely to increase utilization. Pair the chosen method with an AVD scaling plan so hosts can start during demand peaks and be drained and deallocated during quieter periods. Maintain a standardized Siemens NX image - including the validated NX release, AMD GPU driver, required environment variables, and monitoring agents - in Azure Compute Gallery so new hosts can be added consistently. Track active sessions, connection latency, CPU, memory, GPU utilization, GPU memory, storage latency, and NX ATS execution times. Establish scaling thresholds from representative production models rather than using user count alone.

Contributors

Microsoft maintains this article. The following contributors from Microsoft originally wrote it:

Principal author:

• Sunita Phanse | Sr. Technical Program Manager

Contributors:

• Sachin Rastogi | Senior Program Manager

To see nonpublic LinkedIn profiles, sign in to LinkedIn.