Verify STRATUM GPU in the Guest
After Guest Tools are installed, verify that the guest sees the STRATUM GPU device and can query the accelerator before testing CUDA applications.
Verify the device path
Start with STRATUM GPU inventory, then confirm the PCI device and STRATUM device nodes inside the guest. These checks establish that the VM is attached to an accelerator and that Guest Tools are bound to the virtual device.
The physical accelerator can be provided elsewhere in the STRATUM Continuum; the guest consumes the assigned STRATUM GPU without needing to know where that physical device resides.
Interface tour
STRATUM GPU SMI
Run stratum-gpu-smi to confirm that the guest can query the STRATUM-assigned accelerator and see its provider, type, memory, and identity.
STRATUM GPU PCI Device
Confirm the STRATUM GPU PCI device, kernel module, and STRATUM device nodes are present in the guest. These checks verify the device-to-Guest-Tools path.
NVIDIA SMI With STRATUM GPU
Where the supported NVIDIA userspace utility is installed, nvidia-smi can be used as an additional NVML-facing validation of the STRATUM GPU user-mode path.
Commands
Run these checks inside the guest. They verify the supported customer path and intentionally omit development-only library-path, mock-mode, and native-port overrides.
lspci -nn | grep -i -E '3d|display|1d1e:5350'
lsmod | grep stratum_gpu
ls -l /dev/stratum-gpu*
ls -l /dev/stratum-gpuctl
stratum-gpu-query
stratum-gpu-smiThe outcome you should see
The PCI device, STRATUM GPU control/device nodes, inventory query, and GPU status view all agree that the guest has an attached accelerator.