Fast modular exponentiation

[Phil Karn]

An interesting discussion. Thanks.

How about a basic tutorial on these various modexp algorithms, with
particular attention to how many of each arithmetic operation
(add/subtract/multiply/divide) are needed as a function of the modexp
input parameter sizes?  I don't really understand all the details yet,
especially how they relate to which algorithm is best for a given
machine.

I.e., if I come up with a list of clock counts for each basic
arithmetic instruction, how can I tell which algorithm is probably
best for my machine?

Phil