Setting up the emulation layers

Install dependencies

The clone and emulation layer script in this section apply to Unreal Engine 5.4 and 5.6. For Unreal Engine 4.27, use the UE 4.27 plugin branch and follow its README for the matching SDK and installation steps.

To run NSS in your Unreal Engine project, you will need to install and configure the following:

• Vulkan SDK version 1.4.321.0 or later: Required for developing applications that use Vulkan and for accessing Vulkan Configurator. Vulkan Configurator sets up the emulation layers used for running ML extensions for Vulkan workloads.
• ML Emulation Layer for Vulkan: The plugin download script installs the version required by its bundled ArmNG SDK. Its Graph and Tensor layers run ML workloads through Vulkan’s compute backend.
• Arm Neural Graphics Plugin 1.1.1 for Unreal Engine 5.4 and 5.6: Clone the plugin in this section, then run its emulation layer download script.

These components allow you to run NSS in Unreal Engine, using ML emulation layers for Vulkan for development and testing.

Install Vulkan Software Development Kit

Go to the Vulkan SDK landing page and download version 1.4.321.0 or later of the SDK installer for Windows. After you run the installer, continue to the next step.

Clone the Arm Neural Graphics Plugin

Use the Arm Neural Graphics Plugin 1.1.1 source, including its SDK submodules and Git LFS assets, to set up NSS for Unreal Engine.

Arm Neural Graphics Plugin → GitHub repository

With Git and Git LFS installed, run the following in PowerShell to download the source and its dependencies. GitHub’s automatically generated source ZIP omits the submodule contents.

    

        
        
git lfs install
git clone --branch ue5 --recurse-submodules https://github.com/arm/neural-graphics-for-unreal.git
cd neural-graphics-for-unreal
git lfs pull
git submodule foreach --recursive "git lfs pull"

    

The ue5 branch contains NG.uplugin at the repository root.

Download the emulation layers

From the cloned plugin repository root, run:

    

        
        
.\DownloadEmulationLayer.bat

    

The script reads the required version from the bundled ArmNG SDK compatibility matrix and installs these files in Binaries\ThirdParty\VulkanMLEmulation\Win64:

• VkLayer_Graph.dll and VkLayer_Graph.json
• VkLayer_Tensor.dll and VkLayer_Tensor.json

Run the script before configuring Vulkan Configurator. If the download fails, check that Python 3 is on PATH and that GitHub is reachable.

Configure Vulkan Layers

Use Vulkan Configurator to make the downloaded Graph and Tensor layers available to Unreal Engine.

To emulate the ML extensions for Vulkan:

1. Launch the Vulkan Configurator (bundled with the Vulkan SDK) from the Windows Start menu.

2. In the Apply a Vulkan Loader Configuration list, right-click and choose Create a new Configuration. You can give the new configuration any name, for example NSS.

3. Navigate to the Vulkan Layers Location tab.

4. Append a user-defined path pointing to the layers downloaded by the plugin script:

       
   
           
           
   <path-to-neural-graphics-for-unreal>\Binaries\ThirdParty\VulkanMLEmulation\Win64
   
       
   

   Figure 1: Add Vulkan layer path.

5. Confirm that VK_LAYER_ML_Graph_Emulation and VK_LAYER_ML_Tensor_Emulation appear in the layer list.

6. Switch back to the Vulkan Loader Management tab. Set Vulkan Loader Configuration Scope to Any Running Vulkan Executable and select your NSS configuration under Apply a Vulkan Loader Configuration. Select Include Vulkan Layers Selection and Execution Order, set both ML emulation layers to Enable, and place Graph above Tensor.

   Figure 2: Verify layer ordering and scope.

With the ML emulation layers configured, Vulkan can run machine learning workloads through the ML extensions for Vulkan. This enables neural inference to execute alongside the graphics pipeline during development, without requiring access to hardware with dedicated neural accelerators.

The next step is to integrate Neural Super Sampling into an Unreal Engine project. You’ll do this by adding the cloned plugin and creating a simple example game that lets you verify the setup and visualize the upscaling in action.