@John May Thank you for your reply!

May I enquire more how Maple chooses the 'reps' argument? I see one of two options:

• Based on the bit size of the candidate being tested, i.e Damgard, Landrock and Pomerance's Average case error estimates.
• Hard coded to some reasonable amount, say 10, 15 or 25.

Perhaps if you would prefer we could discuss via email. Any evidence would be greatly appreciated.

Thanks again,

Jake