llama.cpp

54739490 - webgpu : Query for adapter support when registering WebGPU backend (#21579)