[x86] Allow (v)divsd->(v)mulsd to execute in parallel.

src/ia32/lithium-codegen-ia32.cc

 // Copyright 2012 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "src/v8.h"
 
 #if V8_TARGET_ARCH_IA32
 
 #include "src/base/bits.h"
 #include "src/code-factory.h"
@@ skipping 2022 matching lines @@
         __ mulsd(left, right);
       }
       break;
     case Token::DIV:
       if (CpuFeatures::IsSupported(AVX)) {
         CpuFeatureScope scope(masm(), AVX);
         __ vdivsd(result, left, right);
       } else {
         DCHECK(result.is(left));
         __ divsd(left, right);
-        // Don't delete this mov. It may improve performance on some CPUs,
-        // when there is a mulsd depending on the result
-        __ movaps(left, left);
       }
+      // Don't delete this mov. It may improve performance on some CPUs,
+      // when there is a (v)mulsd depending on the result
+      __ movaps(result, result);
       break;
     case Token::MOD: {
       // Pass two doubles as arguments on the stack.
       __ PrepareCallCFunction(4, eax);
       __ movsd(Operand(esp, 0 * kDoubleSize), left);
       __ movsd(Operand(esp, 1 * kDoubleSize), right);
       __ CallCFunction(
           ExternalReference::mod_two_doubles_operation(isolate()),
           4);
 
@@ skipping 3705 matching lines @@
   CallRuntime(Runtime::kPushBlockContext, 2, instr);
   RecordSafepoint(Safepoint::kNoLazyDeopt);
 }
 
 
 #undef __
 
 } }  // namespace v8::internal
 
 #endif  // V8_TARGET_ARCH_IA32