WIP: Add first-class support for float32 representations.

src/compiler/ia32/code-generator-ia32.cc

Index: src/compiler/ia32/code-generator-ia32.cc
diff --git a/src/compiler/ia32/code-generator-ia32.cc b/src/compiler/ia32/code-generator-ia32.cc
index c476eaa9eb1aec659be086b7726a8f57a7872bdf..52cd5676fc6f3e03eedf462dd2e85a25909ef100 100644
--- a/src/compiler/ia32/code-generator-ia32.cc
+++ b/src/compiler/ia32/code-generator-ia32.cc
@@ -297,6 +297,14 @@ void CodeGenerator::AssembleArchInstruction(Instruction* instr) {
       __ add(esp, Immediate(kDoubleSize));
       break;
     }
+    case kSSEFloat32ToFloat64:
+      // TODO(turbofan): IA32 SSE conversions should take an operand.
+      __ cvtss2sd(i.OutputDoubleRegister(), i.InputDoubleRegister(0));
+      break;
+    case kSSEFloat64ToFloat32:
+      // TODO(turbofan): IA32 SSE conversions should take an operand.
+      __ cvtsd2ss(i.OutputDoubleRegister(), i.InputDoubleRegister(0));
+      break;
     case kSSEFloat64ToInt32:
       __ cvttsd2si(i.OutputRegister(), i.InputOperand(0));
       break;