ggml : backport ggml-alloc from llama.cpp (#433)

ggml : backport ggml-alloc from llama.cpp (#433)

* ggml: add graph tensor allocator (#2411)

Backport a113689571420fb4d6540f1a324d12965781356a from llama.cpp

* ggml-alloc: Don't try to re-use buffers of external tensors

They might be weights that came from another context, so we
have no control over them (and they might be re-used elsewhere
so writing to them would be a bad idea).

* ggml-alloc: Fix some compile errors when GGML_ALLOCATOR_DEBUG is on

---------

Co-authored-by: slaren <redacted>