[clang] [llvm] [AMDGPU] Add support for `v_tanh_bf16` on gfx1250 (PR #147425)

[Matt Arsenault via llvm-commits]

================
@@ -13658,6 +13658,7 @@ bool SITargetLowering::isCanonicalized(Register Reg, const MachineFunction &MF,
     case Intrinsic::amdgcn_frexp_mant:
     case Intrinsic::amdgcn_fdot2:
     case Intrinsic::amdgcn_trig_preop:
+    case Intrinsic::amdgcn_tanh:
----------------
arsenm wrote:

I don't follow, this shouldn't be hard to test? e.g. fcanonicalize-elimination.ll, you should just need a canonicalize call after these which should fold out 

https://github.com/llvm/llvm-project/pull/147425