Table of Contents

System Requirements

  • Operating System:
    • Windows 10 or later
    • Ubuntu 22.04 LTS
  • CPU Architecture: x64 (ARM not yet supported)
  • Browser: Chrome 115, Edge 115, Firefox 115 or Safari 16, or any later version. The web interface is built to that language level, and an older browser cannot run it. The 3D view is rendered on the server and arrives as a stream of images, so the browser itself needs no graphics capability
  • Runtime: none to install. The web service is published self-contained, so it carries the runtime it needs. A .NET project that references the HiAPI packages directly is the one case that needs a runtime installed, and it must be .NET 10.0 — every published package targets it, and an earlier runtime cannot load them at all
  • Network: the cloud package of HiNC for Windows signs in to its license server over HTTPS and syncs with it about every ten minutes while it runs; without a connection it machines for at most an hour after the last successful sync — see Signing In with a Cloud Account
  • Memory (RAM):
    • Minimum: 8GB RAM (suitable for low-resolution models)
    • Recommended: 128GB RAM or higher (for large and detailed models)
    • What a given project needs is estimated from its NC program and its workpiece before it is built — see Memory Planning
  • Graphics:
    • OpenGL 3.3 compatible graphics card or integrated graphics, with the graphics vendor's driver installed. Without that driver Windows by itself offers only its software OpenGL 1.1, which is too old for the 3D view
    • Most computers manufactured within the last 15 years meet this requirement
  • Windows Remote Desktop: HiNC does not support hardware-accelerated 3D inside a Windows Remote Desktop (RDP) session. A HiNC started inside a Remote Desktop session, or run as a Windows service, may be handed only Microsoft's software OpenGL 1.1, and it then runs without the 3D view. Start HiNC on the computer's own desktop and use it from a browser. The browser has no such limit: it only shows the images the server draws and needs no graphics capability of its own, so it works inside a Remote Desktop session too. To administer that computer remotely, use a remote-access tool that shows the computer's own screen (for example VNC, Parsec, HP Anyware or Amazon DCV) instead of signing in over Remote Desktop. On Windows 10 and 11 a Remote Desktop sign-in takes the desktop HiNC runs on off the computer's own screen, and it stays off after the Remote Desktop window is closed, until someone signs in at the computer again; a HiNC that starts or restarts meanwhile, for example into an update, counts as started inside a Remote Desktop session

Performance Guidelines for High-Resolution Simulation

Based on internal testing and practical experience:

CPU Considerations

  • High clock speed is prioritized over multiple cores
  • Single-thread performance is more important than multi-core parallelization
  • For small NC programs: 16GB RAM is workable

Graphics and GPU

  • GPU is used only for 3D rendering/visualization, not for computation acceleration
  • The rendering engine is initialised at startup even if nobody looks at the 3D view. A host with no usable OpenGL context, or with one older than 3.3 (for example Microsoft's software OpenGL 1.1, which names itself GDI Generic), still starts and logs a warning naming the OpenGL it was given, if any, and the likely cause: NC play, physics and training run, and the 3D view shows a notice saying why it is off instead of a picture
  • On Windows, OpenGL 1.1 usually means that HiNC was started or restarted inside a Remote Desktop session, runs as a Windows service, or found no graphics vendor driver (see Windows Remote Desktop above)
  • Cloud deployment is supported; for the 3D view choose an instance that has a GPU with its graphics driver installed. An instance with no GPU, or with a GPU whose driver is missing or mismatched, runs headless without 3D

Storage Performance

  • SSD (Solid State Drive) recommended for optimal simulation efficiency
  • Traditional HDD may cause significant performance degradation.