llama.cpp

Query for adapter support when registering WebGPU backend

#21579

Merged