MIPS: [turbofan] Introduce optional Float64Min and Float64Max machine operators.

Port 99f8d57f

Original commit message:
Basically recognize certain x < y ? x : y constructs and turn that into
Float64Min/Float64Max operations, if the target machine supports that.
On x86 we lower to (v)minsd/(v)maxsd.

BUG=

Review URL: https://codereview.chromium.org/1006513002

Cr-Commit-Position: refs/heads/master@{#27168}

src/compiler/mips/instruction-selector-mips.cc

@@ -437,6 +437,12 @@ void InstructionSelector::VisitFloat64Mod(Node* node) {
 }
 
 
+void InstructionSelector::VisitFloat64Max(Node* node) { UNREACHABLE(); }
+
+
+void InstructionSelector::VisitFloat64Min(Node* node) { UNREACHABLE(); }
+
+
 void InstructionSelector::VisitFloat64Sqrt(Node* node) {
   MipsOperandGenerator g(this);
   Emit(kMipsSqrtD, g.DefineAsRegister(node), g.UseRegister(node->InputAt(0)));

src/compiler/mips64/instruction-selector-mips64.cc

@@ -617,6 +617,12 @@ void InstructionSelector::VisitFloat64Mod(Node* node) {
 }
 
 
+void InstructionSelector::VisitFloat64Max(Node* node) { UNREACHABLE(); }
+
+
+void InstructionSelector::VisitFloat64Min(Node* node) { UNREACHABLE(); }
+
+
 void InstructionSelector::VisitFloat64Sqrt(Node* node) {
   Mips64OperandGenerator g(this);
   Emit(kMips64SqrtD, g.DefineAsRegister(node), g.UseRegister(node->InputAt(0)));