Use the Remote AI service

Install the Remote AI service on one powerful computer, providing faster and more effective AI capabilities to weaker UFT One computers.

Overview

Install the Remote AI service on a central powerful computer, and connect your UFT One machines remotely to the service. This way, one strong computer provides the AI capabilities for all the UFT One machines.

AI capabilities in UFT One heavily rely on sophisticated computations and involve graphic processing. We strongly recommend that you use a GPU-powered computer for the Remote AI service, to optimize the performance of the AI-based functionality. For specific system requirements, see the Support Matrix.

Back to top

Prerequisites

AI service prerequisites

To be able to use the Remote AI service, make sure you have the following installed on your remote computer:

  • A 64-bit operating system, Windows 10 or later.

  • Microsoft Visual C++ 2013 Redistributable

  • The required Windows feature, depending on your Windows operating system:

    For the most up-to-date instructions on enabling Windows features, see the Microsoft documentation.

GPU prerequisites

To enable GPU processing in the Remote AI service:

  • Make sure your remote computer has a NVIDIA graphics card.

  • Make sure you have the following installed on your remote computer. For installation details, see Install the GPU prerequisites.

    • NVIDIA CUDA Toolkit 11.0 Update 1

    • NVIDIA cuDNN 8.0.2 for CUDA 11.0

Back to top

Install the GPU prerequisites

To enable GPU computing and processing when you are using the Remote AI service in UFT One, you must install the required NVIDIA GPU packages, including the NVIDIA CUDA Toolkit and its corresponding cuDNN library.

To install the NVIDIA CUDA Toolkit for Windows

  1. Verify that you have a CUDA-capable GPU. See Verify you have a CUDA-capable GPU.

  2. Download the NVIDIA CUDA Toolkit 11.0 Update 1 from the download site.

  3. Install the CUDA software by executing the CUDA installer and following the on-screen prompts. Alternatively, you can use the silent installation. For details, see Install the CUDA software.

To install the NVIDIA cuDNN 8.0.2 for CUDA 11.0

  1. Make sure you are registered for the NVIDIA Developer Program.

  2. Download the NVIDIA cuDNN 8.0.2 for CUDA 11.0 from the NVIDIA cuDNN home page.

  3. Follow the instructions described in Installing on Windows in NVIDIA cuDNN documentation.

Back to top

Install the AI service on a remote computer

Install the AI service on a remote computer, which is on the same network as your local UFT One computers.

To install the AI service

  1. Download the AI service installation package from the Micro Focus Marketplace, and extract it.

    Note: The remote AI service version must match the version of your local UFT One.

  2. Install the AI service using a wizard UI to guide you or using a command line.

    As part of the installation, you can provide a port number. This is the port used for communication between the remote AI service and your local UFT One computers. If you do not specify a port number, 443 is used.

    Install the AI service with a wizard

    Run the setup.exe file and follow the steps as instructed to complete your installation.

    Install the AI service silently
    1. Run the following command to install the prerequisites:

      setup.exe /InstallOnlyPrerequisite /s
    2. Run the following command to install the AI service:

      msiexec /i Remote_AI_Installer.msi /qb PORT=<AI service port number>

Back to top

Connect your local UFT One to the Remote AI service

To connect your local UFT One to the Remote AI service, configure the connection information in the Tools > Options pane. For details, see Configure the Remote AI Service connection.

If the connection is successful, click OK in the Tools > Options > AI > Remote AI Service pane, and then UFT One will use the remote service for AI Features instead of the local AI service.

Back to top