or download this
Benchmark: timing 100 iterations of 1_array, 3_array, Baseline, LR_1..
+.
1_array: 10.651 wallclock secs (10.65 usr + 0.00 sys = 10.65 CPU)
+@ 9.39/s (n=100)
3_array: 9.28456 wallclock secs ( 9.28 usr + 0.00 sys = 9.28 CPU)
+ @ 10.78/s (n=100)
...
LR_1 14245/s 1% -- -27% -70%
3_array 19646/s 39% 38% -- -58%
Baseline 47170/s 233% 231% 140% --