Re: Extremely poor performance crunching random numbers under PIV-FC5

[BankHacker <bankhacker gmail com>]

Moreover, I have implemented a good random function I have found here:

http://astronomy.swin.edu.au/~pbourke/other/random/randomlib.c

I have compiled it as a dynamic library, in order to simulate the same
conditions that glibc's rand() function.

I have compiles the test and gives these results:

# ./libcliente
10 M de rand() en 46.950 s. (ejemplo.: 186687031)
10 M de fakerand() en 1.060 s. (ejemplo.: 867)

Recycling code I have called this external function with the old name
"fakerand", but it is a good alternative to standard random
generators.

Conclusion: The system behaves correctly when invoking another random
algorithm dynamically.