llama.cpp GPU Usage Failure Due to Missing Drivers

This cause is identified when the application is built for GPU support but fails to initialize the hardware. You can diagnose this by checking the system logs for missing runtime libraries (such as libcuda.so or cudart) or by confirming that your installed driver version is incompatible with the version of the CUDA Toolkit required by your specific llama.cpp binary.

Sources