Qualify NVIDIA GPUs by per-device CUDA compute capability

Problem

The NVIDIA hardware probe records the global CUDA driver API version from nvmlSystemGetCudaDriverVersion_v2 and copies its major version onto every detected GPU as CudaMajorVersion.

https://github.com/openclaw/openclaw-windows-node/blob/31dffdb0dbd9eebc8b98ce35cf13ab60762410a9/src/OpenClaw.Shared/Inference/NvmlHostHardwareProbe.cs#L174-L207

Eligibility then compares that global driver value with the runtime CUDA version and reports CudaCapabilityTooLow:

https://github.com/openclaw/openclaw-windows-node/blob/31dffdb0dbd9eebc8b98ce35cf13ab60762410a9/src/OpenClaw.Shared/Inference/Catalog/LocalInferenceEligibility.cs#L121-L138

The CUDA driver API version and a GPU's compute capability are different facts. The current code does not call the per-device nvmlDeviceGetCudaComputeCapability API and therefore does not prove that each selected GPU architecture is supported by the pinned llama-server CUDA build.

NVIDIA documents the per-device query here:

https://docs.nvidia.com/deploy/nvml-api/group__nvmlDeviceQueries.html

Impact

A sufficiently new installed driver can satisfy the global CUDA-version check without establishing that a particular GPU is compatible with the architecture set supported by the pinned runtime artifacts. On mixed-generation systems, eligibility can select a device without explicit per-device compatibility evidence and discover the incompatibility only after downloading and launching the runtime.

Expected behavior

• Record the CUDA driver API version separately and name it accordingly.
• Probe compute-capability major and minor for every NVML device.
• Define the minimum or supported compute capabilities for each pinned runtime variant.
• Require both a compatible driver and compatible selected device.
• Fail closed when required per-device capability evidence is unavailable.
• Keep selection and launch pinned to the same stable GPU UUID.

Tests

Add coverage for:

• Compatible driver and compatible GPU
• Compatible driver but unsupported GPU compute capability
• Mixed-generation multiple-GPU selection
• Missing or unsupported NVML compute-capability query
• Clear separation of driver-version and device-capability failure reasons

Tracking requested by @joelagnel.

Start with src/OpenClaw.Shared/Inference/NvmlHostHardwareProbe.cs and LocalInferenceEligibility.cs to trace how driver versions and selected GPU UUIDs are used. Add coverage for the five listed compatibility and failure cases, then verify that driver and per-device capability checks remain separate and selection fails closed when capability evidence is unavailable.