[Ecm-dev] sigma / starting point

Paul Zimmermann Paul.Zimmermann at loria.fr
Ven 28 Avr 16:15:41 CEST 2006


       Willemien,

I now understand what happens. Please forget my previous mails.

You first call ecm with B1=260000000 and interrupt it (with Ctrl-C)
at B1=6647107. All primes <= 6647107 were treated for a target B1
of 260000000, but this information is lost in the save file, which
only contains B1=6647107.

Consider for example p=17. For B1=260000000 it is treated with exponent
6 since 17^6 = 24137569 <= 260000000 < 17^7.

But when you restart, the program considers that the effective B1 was only
6647107, i.e. if 17 were treated with exponent 5 only. Therefore, it performs
one more point multiplication by 17, which results in a total exponent of 7
instead of 6.

This explains why you get different values of X in the save files.

Note: this can happen only when the program is interrupted. If you first
call it with B1=6647107 and -save, then resume it with B1=260000000, you 
will get the same X-values.

Paul




Plus d'informations sur la liste de diffusion Ecm-discuss