Optimize smi multiply by 2 using shl by 1.

runtime/vm/intermediate_language_ia32.cc

Index: runtime/vm/intermediate_language_ia32.cc
===================================================================
--- runtime/vm/intermediate_language_ia32.cc	(revision 20202)
+++ runtime/vm/intermediate_language_ia32.cc	(working copy)
@@ -2209,7 +2209,11 @@
       case Token::kMUL: {
         // Keep left value tagged and untag right value.
         const intptr_t value = Smi::Cast(constant).Value();
-        __ imull(left, Immediate(value));
+        if (value == 2) {

[bakster, 2013/03/19 13:22:44]

What about 4, 8, 16 etc?

[Florian Schneider, 2013/03/19 13:24:33]

shl does not set the overflow flag for other than 1-bit shifts.

[bakster, 2013/03/19 13:31:25]

I guess you than can unroll into repeated shl(1) :-)

[ahe, 2013/03/19 14:13:11]

What about a compare?

+          __ shll(left, Immediate(1));
+        } else {
+          __ imull(left, Immediate(value));
+        }
         if (deopt != NULL) __ j(OVERFLOW, deopt);
         break;
       }