Pascal-Architecture GPU (e.g., Quadro P2000) Not Supported for Deep Learning in EthoVision XT 19
Last updated: Oct 29, 2026
Problem
Users upgrading to EthoVision XT 19 receive an error stating 'The graphics card (GPU) is not supported. Please install one with compute capability...' even after updating NVIDIA drivers or installing a newer CUDA toolkit. The issue persists regardless of driver version.
Details
EthoVision XT 19 dropped support for GPUs based on Pascal architecture (compute capability < 7.5). Affected cards include the NVIDIA Quadro P2000 and other Pascal-generation GPUs. Updating drivers (e.g., R580 U10 / 582.70 WHQL) or installing a newer CUDA toolkit (e.g., 13.3) does not resolve the error because the hardware itself does not meet the minimum compute capability requirement of 7.5. Users may also encounter this when checking compatibility before upgrading existing lab computers.
Solution
1. Identify your GPU generation: Pascal-architecture cards (e.g., Quadro P2000, GTX 10-series) have compute capability below 7.5 and are not supported for deep-learning features in EthoVision XT 19. Updating drivers or CUDA will not help.
2. Check Noldus's published minimum GPU specifications for EthoVision XT 19 (minimum compute capability 7.5, e.g., Turing/Ampere-generation cards such as RTX series).
3. If you cannot replace the GPU, switch EthoVision to CPU-based video processing as a workaround for non-deep-learning tasks (see the KB article 'GPU Rendering Issues in EthoVision XT – Switching to CPU Video Processing').
4. For deep-learning features (e.g., automatic behavior recognition), a GPU upgrade to a supported card (compute capability ≥ 7.5) is required.
5. Refer to the Noldus GPU deprecation notice for the full list of dropped GPU architectures and recommended replacement hardware.