Add support AMD gfx1103 780M APU GPU and HIP_UMA

[atelepov]

Support HIP_UMA was added in llama.cpp PR:
ggml-org/llama.cpp#7414

Support ROCm can be implemented by example:
https://github.com/likelovewant/ROCmLibs-for-gfx1103-AMD780M-APU

This all can speedup working with llm on ROCm and AMD 780M APU GPU

[DarkGhostHunter]

Second to this. I believe this GPU can be a good start for supporting the 880M/890M GPU (gfx1151) as well, since both share the same architecture.