[seqfan] Re: Computing sequences up to first non-distinct element: re-inventing the wheel?

[Ron Hardin]

Not Mathematica but what I did for Collatz sequences was run two, one at half 
speed, and check for equality.

It solves the memory problem at some cost in speed.

In the case of Collatz, you only have to start running two if you've broken the 
previous record for number of steps, getting most of the speed back.

 rhhardin at mindspring.com
rhhardin at att.net (either)