System Requirements#

  • Operating System

    • Ubuntu 24.04

    • Windows supported via WSL

  • Driver and GPU Requirements

    • pip: NVIDIA driver that is compatible with local PyTorch installation.

    • Docker container: PhysicsNeMo container release 26.03 is based on PyTorch 26.01 container. For driver requirements, please refer PyTorch NVIDIA Container Release Notes.

  • Required installations for pip install

    • Python >= 3.10

  • Recommended Hardware

    • 64-bit x86

    • NVIDIA GPUs:

      • Blackwell GPUs - B100, RTX PRO 6000 Blackwell

      • Ada Lovelace GPUs - RTX 40xx

      • Hopper GPUs - H100, H200

      • NVIDIA Ampere GPUs - A100, A30, A40, A4000, A6000, RTX 30xx

      • Turing GPUs - T4

    • Other Supported GPUs:

      • NVIDIA Ampere GPUs - RTX 30xx

    • For others, please reach us out at GitHub

    • 64-bit ARM architecture is also supported.

Note

The recommended hardware is based on testing done by the PhysicsNeMo team. You can refer to the Frameworks Support Matrix for a broader set of supported systems where you can run PhysicsNeMo.

Note

To get the benefits of all the performance improvements, use the NVIDIA container for PhysicsNeMo. This container comes with all the prerequisites and dependencies and allows you to get started efficiently with PhysicsNeMo.