[llvm] [AMDGPU] Use v_pk_add_bf16 for scalar bf16 fadd on gfx1250/gfx13 (PR #209128)

[Matt Arsenault via llvm-commits]

https://github.com/arsenm commented:

This should be performed in the legalizer by introducing a vector op with a poison high half. I don't know why V_MAX_BF16_PSEUDO is here, but this should not follow that example 

https://github.com/llvm/llvm-project/pull/209128