Yeah, these days, any standard random number source on a computer should be good enough.
Back in my undergrad days, I wrote up a large-scale die roller for testing probability. I used the school's Sun workstations to do so, and I discovered a strangeness with Sun's random number generator:
If the die type was divisible by 8, the output was *exactly* evenly distributed. (this was for a homebrew chargen system that involved rolling ~40-50 d8) I can only assume there was some incrementing going on with the random seed, as well as a mod 8 calculation.