mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Software (https://www.mersenneforum.org/forumdisplay.php?f=10)
-   -   Spooky sigma values (https://www.mersenneforum.org/showthread.php?t=13773)

lavalamp 2010-08-24 12:20

Spooky sigma values
 
1 Attachment(s)
I'm running two instances of Prime95, each of which is running stage 1 ECM curves, and I've noticed that the sigma values for the curves within each instance is very similar. Specifically, they usually share the first 6 or 7 most significant digits.

I've attached a screen shot so you can see what I mean. Is this deliberate, or does P95 just have problems running curves on the same number with multiple workers with this as a symptom?

Prime95 2010-08-24 14:52

Nothing spooky here, just a byproduct of P95's "random" number generator. IIRC, the upper bits are based on the date, the low bits are based on the timestamp counter.

Obviously, this does not generate random numbers, but it will generate unique numbers, which is the goal.

lavalamp 2010-08-24 15:22

I had wondered if it was based on a timestamp, but then as the start times for the curves drifted apart, the sigma values still lined up, even now when there's a 35 min gap between start times.

Well, however it get's the numbers, I'm glad it's the expected behaviour and that they will be unique. Thank-you for the reassurance.


All times are UTC. The time now is 14:56.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.