[turbofan] Add Float32(Max|Min) machine operators.

src/compiler/machine-operator.cc

Index: src/compiler/machine-operator.cc
diff --git a/src/compiler/machine-operator.cc b/src/compiler/machine-operator.cc
index f2b540665b319bb3624deb9de51bb94851ad682e..459d18671904b4112f249ca733e294f64e67e71f 100644
--- a/src/compiler/machine-operator.cc
+++ b/src/compiler/machine-operator.cc
@@ -164,6 +164,8 @@ MachineRepresentation AtomicStoreRepresentationOf(Operator const* op) {
   V(Float32Div, Operator::kNoProperties, 2, 0, 1)                            \
   V(Float32Neg, Operator::kNoProperties, 1, 0, 1)                            \
   V(Float32Sqrt, Operator::kNoProperties, 1, 0, 1)                           \
+  V(Float32Max, Operator::kAssociative | Operator::kCommutative, 2, 0, 1)    \

[titzer, 2016/08/17 11:26:09]

Are we sure these are associative?

[jbramley, 2016/08/17 12:02:57]

They are associative, but they are only commutative if we don't care about which
NaN is propagated in the case where there are two NaNs.

Also note that since #37971, Float64(Min|Max) should share these properties too
(though that isn't related to this patch).

+  V(Float32Min, Operator::kAssociative | Operator::kCommutative, 2, 0, 1)    \
   V(Float64Abs, Operator::kNoProperties, 1, 0, 1)                            \
   V(Float64Acos, Operator::kNoProperties, 1, 0, 1)                           \
   V(Float64Acosh, Operator::kNoProperties, 1, 0, 1)                          \