VmWare Grid a vs q VGPU profiles for headless?


I’m running a headless AI server with tensor-flow, pytorch, YOLO, CUDA 12.3 and CUDNN on a ESXI VM Windows Server 2022 with the Nvidia grid drivers. Does it make any difference whether I chose the nvidia_rtxa5000-8a or the nvidia_rtxa5000-8q profile? Wasn’t able to determine this from the docs.

best regards

vApps doesn’t support CUDA so you will need to use a Q profile for this use case.