================
@@ -214,13 +223,41 @@ struct APFloatBase {
// types, there are no infinity or NaN values. The format is detailed in
//
https://www.opencompute.org/documents/ocp-microscaling-formats-mx-v1-0-spec-final-pdf
S_Float4E2M1FN,
-
+ // TODO: Documentation is missing.
S_x87DoubleExtended,
- S_MaxSemantics = S_x87DoubleExtended,
+ // These are legacy semantics for the fallback, inaccrurate implementation
+ // of IBM double-double, if the accurate semPPCDoubleDouble doesn't handle
+ // the operation. It's equivalent to having an IEEE number with consecutive
+ // 106 bits of mantissa and 11 bits of exponent.
+ //
+ // It's not equivalent to IBM double-double. For example, a legit IBM
+ // double-double, 1 + epsilon:
+ //
+ // 1 + epsilon = 1 + (1 >> 1076)
+ //
+ // is not representable by a consecutive 106 bits of mantissa.
+ //
+ // Currently, these semantics are used in the following way:
+ //
+ // semPPCDoubleDouble -> (IEEEdouble, IEEEdouble) ->
+ // (64-bit APInt, 64-bit APInt) -> (128-bit APInt) ->
+ // semPPCDoubleDoubleLegacy -> IEEE operations
+ //
+ // We use bitcastToAPInt() to get the bit representation (in APInt) of the
+ // underlying IEEEdouble, then use the APInt constructor to construct the
+ // legacy IEEE float.
+ //
+ // TODO: Implement all operations in semPPCDoubleDouble, and delete these
+ // semantics.
+ S_PPCDoubleDoubleLegacy,
----------------
arsenm wrote:
So it seems like this was intentionally not added here as a publicly visible
kind?
https://github.com/llvm/llvm-project/pull/117291
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
