https://github.com/doru1004 requested changes to this pull request. Can you provide a similar intrinsic which can be called as Intrinsic::int_amdgcn_global_atomic_fmax which calls your desired intrinsic instead? https://github.com/llvm/llvm-project/pull/105642