llama.cpp

3e941b81 - ggml: update SCHED_DEBUG output to use ggml_op_desc() (#22825)