Benchmark: timing 25000 iterations of cclever3, clever2, clever3, csimple3, simple3... cclever3: 1 wallclock secs ( 0.95 usr + 0.00 sys = 0.95 CPU) @ 26315.79/s (n=25000) clever2: 10 wallclock secs ( 8.73 usr + 0.02 sys = 8.75 CPU) @ 2857.14/s (n=25000) clever3: 1 wallclock secs ( 1.15 usr + 0.00 sys = 1.15 CPU) @ 21739.13/s (n=25000) csimple3: 1 wallclock secs ( 1.04 usr + -0.01 sys = 1.03 CPU) @ 24271.84/s (n=25000) simple3: 5 wallclock secs ( 3.97 usr + 0.01 sys = 3.98 CPU) @ 6281.41/s (n=25000)