George responded:
>
> At 07:49 PM 11/3/98 +0100, you wrote:
> >
> >Does that take into account the fact that a factor found does not
> >require time to check, whereas a LL residue needs to be
> >computed twice?
>
> Yes and no. I don't take it into account when I compute
> the breakeven point, but I always use the best case factoring code.
> Thus I compute the limit using PPro and PII factoring timings which are
> twice as fast as a Pentium of similar clock speed. Thus Pentiums
> are probably over-factoring and PPros and PIIs are under-factoring.
>
> Regards,
> George
