Comparing large arrays

I have two arrays, generated from even larger text source files (one is 
~50K elements and the other is about 90K elements).  I need to compare the 
two, and produce a result array that contains only the elements of array 1 
that are not also in array 2.  It seems like array_diff() is exactly the 
tool needed, but it times out.  So my questions are -

1.  Should I be getting a time out with this size job (I don't have a good 
feel for this - although 5x10^9 comparisons certainly sounds like lots!)?
2.  Is there a better way to do this?

Thanks,

--

Murray