[x64] Use Cvttsd2siq in TruncateDoubleToI for better performance.

R=verwaest@chromium.org, bmeurer@chromium.org
BUG=chromium:603470
LOG=N

Review URL: https://codereview.chromium.org/1890813002

Cr-Commit-Position: refs/heads/master@{#35489}

src/compiler/x64/code-generator-x64.cc

@@ -800,10 +800,13 @@ void CodeGenerator::AssembleArchInstruction(Instruction* instr) {
       auto result = i.OutputRegister();
       auto input = i.InputDoubleRegister(0);
       auto ool = new (zone()) OutOfLineTruncateDoubleToI(this, result, input);
-      __ Cvttsd2si(result, input);
-      __ cmpl(result, Immediate(1));
+      // We use Cvttsd2siq instead of Cvttsd2si due to performance reasons. The
+      // use of Cvttsd2siq requires the movl below to avoid sign extension.
+      __ Cvttsd2siq(result, input);
+      __ cmpq(result, Immediate(1));
       __ j(overflow, ool->entry());
       __ bind(ool->exit());
+      __ movl(result, result);
       break;
     }
     case kArchStoreWithWriteBarrier: {

src/x64/macro-assembler-x64.cc

@@ -3744,8 +3744,8 @@ void MacroAssembler::TruncateHeapNumberToI(Register result_reg,
                                            Register input_reg) {
   Label done;
   Movsd(xmm0, FieldOperand(input_reg, HeapNumber::kValueOffset));
-  Cvttsd2si(result_reg, xmm0);
-  cmpl(result_reg, Immediate(1));
+  Cvttsd2siq(result_reg, xmm0);
+  cmpq(result_reg, Immediate(1));
   j(no_overflow, &done, Label::kNear);
 
   // Slow case.
@@ -3767,8 +3767,8 @@ void MacroAssembler::TruncateHeapNumberToI(Register result_reg,
 void MacroAssembler::TruncateDoubleToI(Register result_reg,
                                        XMMRegister input_reg) {
   Label done;
-  Cvttsd2si(result_reg, input_reg);
-  cmpl(result_reg, Immediate(1));
+  Cvttsd2siq(result_reg, input_reg);
+  cmpq(result_reg, Immediate(1));
   j(no_overflow, &done, Label::kNear);
 
   subp(rsp, Immediate(kDoubleSize));