[llvm] [APFloat] Add APFloat support for E8M0 type (PR #107127)

[Durgadoss R via llvm-commits]

================
@@ -955,6 +965,12 @@ void IEEEFloat::makeNaN(bool SNaN, bool Negative, const APInt *fill) {
       significand[part] = 0;
   }
 
+  // For the E8M0 types, precision is just 1 and the
+  // the NaNBit handling below is not relevant.
----------------
durga4github wrote:

Cleaned-up in the latest revision,

https://github.com/llvm/llvm-project/pull/107127