LangRef: Clarify expected sNaN behavior for minnum/maxnum

[arsenm]

Summary

This matches the de-facto behavior based on constant folding
and the default lowering to fmin/fmax.

[hfinkel]

docs/LangRef.rst

11486

I don't understand this text because it seems like if the standard says that a QNaN is returned if either input is a SNaN, then why would the lowering need to additionally canonicalize the inputs? It seems like that would not change the result.

[arsenm]

docs/LangRef.rst

11486

The standard says minnum(x, snan) -> qnan. This differs from the behavior from fmin, where fmin(x, snan) -> x. You need to quiet to get the other operand instead of qnan

[arsenm]

Try to be clearer.

One thing I am still unclear ones is if the standard requires payload bits to be zero or preserved

[arsenm]

ping

[arsenm]

ping

[arsenm]

ping

[arsenm]

ping

[hfinkel]

In D49978#1206026, @arsenm wrote:

ping

LGTM

[arsenm]

r340762

docs/LangRef.rst

 Semantics:
 """"""""""
 
-Follows the IEEE-754 semantics for minNum, which also match for libm's
-fmin.
+Follows the IEEE-754 semantics for minNum, except for handling of
+signaling NaNs. This match's the behavior of libm's fmin.
 
 If either operand is a NaN, returns the other non-NaN operand. Returns
-NaN only if both operands are NaN. If the operands compare equal,
-returns a value that compares equal to both operands. This means that
-fmin(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
+NaN only if both operands are NaN. The returned NaN is always
+quiet. If the operands compare equal, returns a value that compares
+equal to both operands. This means that fmin(+/-0.0, +/-0.0) could
+return either -0.0 or 0.0.
+
+Unlike the IEEE-754 2008 behavior, this does not distinguish between
+signaling and quiet NaN inputs. If a target's implementation follows
+the standard and returns a quiet NaN if either input is a signaling
+NaN, the intrinsic lowering is responsible for quieting the inputs to
+correctly return the non-NaN input (e.g. by using the equivalent of
+``llvm.canonicalize``).
+
 
 '``llvm.maxnum.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

 
 Semantics:
 """"""""""
-Follows the IEEE-754 semantics for maxNum, which also match for libm's
-fmax.
+Follows the IEEE-754 semantics for maxNum except for the handling of
+signaling NaNs. This matches the behavior of libm's fmax.
 
 If either operand is a NaN, returns the other non-NaN operand. Returns
-NaN only if both operands are NaN. If the operands compare equal,
-returns a value that compares equal to both operands. This means that
-fmax(+/-0.0, +/-0.0) could return either -0.0 or 0.0.
+NaN only if both operands are NaN. The returned NaN is always
+quiet. If the operands compare equal, returns a value that compares
+equal to both operands. This means that fmax(+/-0.0, +/-0.0) could
+return either -0.0 or 0.0.
+
+Unlike the IEEE-754 2008 behavior, this does not distinguish between
+signaling and quiet NaN inputs. If a target's implementation follows
+the standard and returns a quiet NaN if either input is a signaling
+NaN, the intrinsic lowering is responsible for quieting the inputs to
+correctly return the non-NaN input (e.g. by using the equivalent of
+``llvm.canonicalize``).
 
 '``llvm.copysign.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^