Most NVIDIA GPUs can be configured in either WDDM or TCC mode.
NVIDIA GPUs based on the Tesla architecture default to TCC mode whereas those based on previous architectures like Pascal default to WDDM mode. The mode can be switched on Telsa and Pascal based cards using the nvidia-smi utility:
https://docs.nvidia.com/gameworks/content/developertools/desktop/nsight/tesla_compute_cluster.htm
The hardware acceleration section of your requirements page does not address the GPU mode. It only states:
https://www.milestonesys.com/systemrequirements/
Decoding with NVIDIA graphics card is supported with GPU capability version 6.x (Pascal) or newer.
Should the GPU in the recording server be in WDDM or TCC mode when using server side motion detection?
Specifically, if we use an NVIDIA T4 which default to TCC mode, do we need to switch it to WDDM mode for server side motion detection?