Improved method invocation performance: 1.5x for virtual and 2.8x for interface.

- Implemented predicted chaining for invoke virtual and interface.
- Eliminated a little bit of fat for invoke native.
- Added 078-polymorphic-virtual for stress tests.
diff --git a/vm/compiler/CompilerIR.h b/vm/compiler/CompilerIR.h
index 8a2028a..201b0cc 100644
--- a/vm/compiler/CompilerIR.h
+++ b/vm/compiler/CompilerIR.h
@@ -23,7 +23,8 @@
     /* For coding convenience reasons chaining cell types should appear first */
     CHAINING_CELL_NORMAL = 0,
     CHAINING_CELL_HOT,
-    CHAINING_CELL_INVOKE,
+    CHAINING_CELL_INVOKE_SINGLETON,
+    CHAINING_CELL_INVOKE_PREDICTED,
     CHAINING_CELL_LAST,
     DALVIK_BYTECODE,
     PC_RECONSTRUCTION,