[llvm] [AMDGPU] Disable inline constants for pseudo scalar transcendentals (PR #104395)

[Carl Ritson via llvm-commits]

================
@@ -5776,6 +5776,18 @@ bool SIInstrInfo::isOperandLegal(const MachineInstr &MI, unsigned OpIdx,
           return false;
       }
     }
+  } else if (isVOP3(MI) && ST.hasNoF16PseudoScalarTransInlineConstants() &&
----------------
perlfu wrote:

Agreed, the subtarget test will be slightly faster than the MI descriptor fetch and bit test.

https://github.com/llvm/llvm-project/pull/104395