llama.cpp

docs: update CPU backend ops to mark POOL_1D as supported

#20304

Merged