ayokunle321 wrote: Hi @skc7,
When implementing __builtin_amdgcn_tanhf I got an error saying it needs the tanh-insts target feature — looks like this is only available on newer targets (gfx950+?). For the other builtins the tests are split across files by type (vi for half, generic for f32/f64, gfx1250 for bfloat16). Should the tanh tests follow that same split, or should they all live in the gfx1250 file since the feature is gated? https://github.com/llvm/llvm-project/pull/197852 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
