Vulkan support on enterprise hardware

Hi,
we use Vulkan API to do some compute and rasterization tasks in headless mode, without ever presenting to screen, and we really admire the quality of Vulkan implementation on GTX/RTX/Quadro cards, but were not able to run Vulkan on Tesla cards (T4, M60, both Windows and Linux). As I see, the support is claimed in the docs here Virtual GPU Software User Guide :: NVIDIA Virtual GPU Software Documentation and here Version 410.72(Linux)/411.82(Windows) :: NVIDIA Tesla Documentation, but when initializing the vulkan instance, we get vk_error_initialization_failed. Vulkan Caps Viewer also reports that it cannot enumerate devices. So,

  1. Is it possible to run Vulkan rendering on M60/T4/V100 cards?
  2. Is it possible to run Vulkan in virtual environment, specifically on Amazon EC2 instances?
  3. Do you have any manuals on this topic/plans to support?

For anyone finding this in the future, I was able to solve a VK_ERROR_INITIALIZATION_FAILED (code -3) error by installing libnvidia-gl-570 in the container using nvidia/cuda:12.8.0-base-ubuntu24.04 as base .