llama.cpp

64b38b56 - server: skip device enumeration in router mode to avoid creating CUDA primary context (#23137)