llama.cpp

a8cbab20 - ggml: add `GGML_SET` Metal kernel + i32 CPU kernel (ggml/1037)

ggml: add `GGML_SET` Metal kernel + i32 CPU kernel (ggml/1037) * implemented cpu kernel * add i32 test cases in test-backend-ops * typedef `ggml_metal_kargs_set` * implemented `kernel_set` * memcpy