New instruction simplifications. Extra dce pass. Allow more per block repeats.

Rationale:
We were missing some obvious simplifications, which left performance
at the table for e.g. CaffeineLogic compiled with dx (4200us->2700us).
The constant for allowing a repeat on a BB seemed very low, at the
very least it should depend on the BB size.

Test: test-art-host

Change-Id: Ic234566e117593e12c936d556222e4cd4f928105
diff --git a/test/611-checker-simplify-if/src/Main.java b/test/611-checker-simplify-if/src/Main.java
index 7dac007..774f239 100644
--- a/test/611-checker-simplify-if/src/Main.java
+++ b/test/611-checker-simplify-if/src/Main.java
@@ -64,13 +64,13 @@
 
   // Test when the phi is the input of the if.
 
-  /// CHECK-START: void Main.testInline(java.lang.String[]) dead_code_elimination$final (before)
+  /// CHECK-START: void Main.testInline(java.lang.String[]) dead_code_elimination$after_inlining (before)
   /// CHECK-DAG: <<Const0:i\d+>>   IntConstant 0
   /// CHECK-DAG:                   If
   /// CHECK-DAG: <<Phi:i\d+>>      Phi
   /// CHECK-DAG:                   If [<<Phi>>]
 
-  /// CHECK-START: void Main.testInline(java.lang.String[]) dead_code_elimination$final (after)
+  /// CHECK-START: void Main.testInline(java.lang.String[]) dead_code_elimination$after_inlining (after)
   /// CHECK:      If
   /// CHECK-NOT:  Phi
   /// CHECK-NOT:  If