Unless I'm mistaken, or leaving out a bigger picture, qwerty is saying that your use of $rand is throwing off the benchmark completly.
Random number generation is so slow, you could probably find the $min of 10 fixed values in less time than it takes to generate one $rand value... and certainly less time than it takes to generate 10 of them. So it's no wonder that your results look identical for both functions, you are benchmarking $rand not $min.
I could also bring up the time you used $lines in a WHILE loop, but that would be hitting below the belt.
What is sexy in C is not sexy in mIRC. -- A mighty scripter.
