$ perl sortbench.pl
Benchmark: timing 1000 iterations of f_alpha, f_owtdi, f_sship, i_alpha,
 i_owtdi, i_sship...
   f_alpha: 57 wallclock secs (51.10 usr +  0.02 sys = 51.12 CPU) @ 19.56/s (n=1000)
   f_owtdi: 92 wallclock secs (83.53 usr +  0.01 sys = 83.54 CPU) @ 11.97/s (n=1000)
   f_sship: 37 wallclock secs (33.55 usr +  0.00 sys = 33.55 CPU) @ 29.81/s (n=1000)
   i_alpha: 41 wallclock secs (36.44 usr +  0.01 sys = 36.45 CPU) @ 27.43/s (n=1000)
   i_owtdi: 91 wallclock secs (82.73 usr +  0.02 sys = 82.75 CPU) @ 12.08/s (n=1000)
   i_sship: 38 wallclock secs (32.81 usr +  0.00 sys = 32.81 CPU) @ 30.48/s (n=1000)
          Rate f_owtdi i_owtdi f_alpha i_alpha f_sship i_sship
f_owtdi 12.0/s      --     -1%    -39%    -56%    -60%    -61%
i_owtdi 12.1/s      1%      --    -38%    -56%    -59%    -60%
f_alpha 19.6/s     63%     62%      --    -29%    -34%    -36%
i_alpha 27.4/s    129%    127%     40%      --     -8%    -10%
f_sship 29.8/s    149%    147%     52%      9%      --     -2%
i_sship 30.5/s    155%    152%     56%     11%      2%      --