ggml : drop support for QK_K=64 (llama/7473)

ggml-opencl, llama: using reserve() if count already known (llama/7272)

opencl : alignment size converted from bits to bytes (llama/7090)

sync : ggml (#2001)

ggml : introduce ggml_status (ggml/750)

code : normalize enum names (llama/5697)

ggml : add max buffer sizes to opencl and metal backends (llama/5181)

ggml : add Vulkan backend (llama/2059)

Add OpenCL add kernel (llama/5151)

llama : ggml-backend integration (llama/4766)

sync : ggml (new ops, new backend, etc) (#1602)

sync : ggml (backend v2, k-quants, CUDA opts, Metal opts, etc.) (#1422)

ggml : sync latest llama.cpp (view_src + alloc improvements) (#1247)

ggml : sync (ggml-alloc, GPU, eps, etc.) (#1220)

opencl : sync opencl compilation fix in ggml (#1111)

ggml : sync latest repo (mostly refactoring changes)

opencl : sync latest ggml-opencl