or download this
Benchmark: timing 10000 iterations of ctypeFor, foreachglobal, foreach
+my, foreachnovar, fornovar...
...
foreachnovar: 23 wallclock secs (23.33 usr + 0.00 sys = 23.33 CPU)
fornovar: 24 wallclock secs (23.44 usr + 0.03 sys = 23.47 CPU)