ARM: Add new String.compareTo intrinsic.

Benchmarked on Nexus6P big, little, and all cores. The new intrinsic is
faster than pStringCompareTo for compare lengths on [1,512], so the
runtime call is no longer needed.

Change-Id: If853b592dfc5e561ea3389b51729f37a2c89c18e
diff --git a/runtime/arch/stub_test.cc b/runtime/arch/stub_test.cc
index 02629e8..a7d6d6f 100644
--- a/runtime/arch/stub_test.cc
+++ b/runtime/arch/stub_test.cc
@@ -1205,9 +1205,9 @@
 
 
 TEST_F(StubTest, StringCompareTo) {
-  // There is no StringCompareTo runtime entrypoint for __aarch64__.
-#if defined(__i386__) || defined(__arm__) || \
-    defined(__mips__) || (defined(__x86_64__) && !defined(__APPLE__))
+  // There is no StringCompareTo runtime entrypoint for __arm__ or __aarch64__.
+#if defined(__i386__) || defined(__mips__) || \
+    (defined(__x86_64__) && !defined(__APPLE__))
   // TODO: Check the "Unresolved" allocation stubs
 
   Thread* self = Thread::Current();