ECM found factors, why so many repeats?

Alles zum Projekt yoyo@home
Everything about the project yoyo@home

ECM found factors, why so many repeats?

Unread postby Tom Poleski » 03.12.2018 08:44

I'm very curious to know why, at the page: http://www.rechenkraft.net/yoyo//y_factors_ecm.php it repeats the same factors found, over and over again. Why is this happening?
Tom Poleski
Idle-Sammler
Idle-Sammler
 
Posts: 3
Joined: 03.12.2018 08:36

Re: ECM found factors, why so many repeats?

Unread postby yoyo » 03.12.2018 22:09

To find a factor thousands of workunits have to be computed. This is done in parallel by hundreds of user. So if an easy factor is found it is possible, that the factor is found by many user in parallel. As fast as the server gets a found factor it cancels all not yet sent out workunits and if a client is connected it tells him to abort all not yet started workunits.
HILF mit im Rechenkraft-WiKi, dies gibts zu tun.
Wiki - FAQ - Verein - Chat

Image Image
User avatar
yoyo
Vereinsvorstand
Vereinsvorstand
 
Posts: 7653
Joined: 17.12.2002 14:09
Location: Berlin

Re: ECM found factors, why so many repeats?

Unread postby Tom Poleski » 04.12.2018 10:48

yoyo wrote:To find a factor thousands of workunits have to be computed. This is done in parallel by hundreds of user. So if an easy factor is found it is possible, that the factor is found by many user in parallel. As fast as the server gets a found factor it cancels all not yet sent out workunits and if a client is connected it tells him to abort all not yet started workunits.


But, as far as I understand, each workunit should be attempting to compute a different range, no? Otherwise would it not be a waste of resources? If multiple computers find the same factor in parallel, doesn't that mean that multiple computers are computing the same workunit?? Or at least, performing the same calculation on the same range of data across multiple work units?
Perhaps I just do not understand enough about how the mathematics works behind it. Thank you for the reply.
Tom Poleski
Idle-Sammler
Idle-Sammler
 
Posts: 3
Joined: 03.12.2018 08:36

Re: ECM found factors, why so many repeats?

Unread postby yoyo » 04.12.2018 20:32

Workunits are NOT splited based on different ranges!
ECM, the eliptic curve method, is a trial factoring method. For a batch we have to run different trials, all starting with a different random seed.
So all computations are different, they are different trials. For "easy" factors it is therefore possible that with different random seeds the same factor is found.
HILF mit im Rechenkraft-WiKi, dies gibts zu tun.
Wiki - FAQ - Verein - Chat

Image Image
User avatar
yoyo
Vereinsvorstand
Vereinsvorstand
 
Posts: 7653
Joined: 17.12.2002 14:09
Location: Berlin

Re: ECM found factors, why so many repeats?

Unread postby Tom Poleski » 06.12.2018 04:12

yoyo wrote:Workunits are NOT splited based on different ranges!
ECM, the eliptic curve method, is a trial factoring method. For a batch we have to run different trials, all starting with a different random seed.
So all computations are different, they are different trials. For "easy" factors it is therefore possible that with different random seeds the same factor is found.


I understand, thank you.
May I just ask though.. what makes a factor "easy"?
Tom Poleski
Idle-Sammler
Idle-Sammler
 
Posts: 3
Joined: 03.12.2018 08:36

Re: ECM found factors, why so many repeats?

Unread postby yoyo » 06.12.2018 06:14

A small factor, below 30 or 40 digits.
HILF mit im Rechenkraft-WiKi, dies gibts zu tun.
Wiki - FAQ - Verein - Chat

Image Image
User avatar
yoyo
Vereinsvorstand
Vereinsvorstand
 
Posts: 7653
Joined: 17.12.2002 14:09
Location: Berlin


Return to Number crunching

Who is online

Users browsing this forum: No registered users and 9 guests