Knowledge Base

Problem

Users with insufficient local GPU resources want to run EthoVision XT 19 Deep Learning analysis on a cloud-based GPU or virtual machine (VM) to avoid purchasing expensive new hardware.

Details

Applies to EthoVision XT 19 with the Automatic Behavior Recognition or body-point deep learning modules. Users whose current computers lack a supported NVIDIA GPU (Turing architecture or newer) explore cloud GPU instances or university HPC VMs as a cost alternative. EthoVision does not officially support cloud or VM deployments.

Solution

1. **Cloud computing and VMs are not supported** configurations for EthoVision XT 19. Noldus cannot guarantee correct behavior, licensing, or performance in these environments.
2. **Recommended local GPU path:**
- EthoVision XT 19 requires an NVIDIA GPU with Turing architecture or newer (e.g., RTX 20-series and above). Pascal-architecture cards (e.g., Quadro P2000) are not supported for deep learning.
- NVIDIA RTX 50-series (e.g., RTX 5060) GPUs with a compatible driver set are expected to work with the latest EthoVision XT 19 updates—check the Noldus release notes for confirmed supported driver versions before purchasing.
- Consult the EthoVision Hardware System Requirements KB article for the full validated GPU and system specification.
3. **Cost alternatives to consider:**
- Purchase a validated Noldus workstation that is pre-configured and tested with EthoVision XT 19.
- Add a supported GPU card to an existing compatible PC if the motherboard and power supply allow it.
4. Contact Noldus Support for the current list of tested and recommended GPU models before purchasing hardware.