mersenneforum.org (https://www.mersenneforum.org/index.php)
-   No Prime Left Behind (https://www.mersenneforum.org/forumdisplay.php?f=82)
-   -   Team drive #8 k=1400-2000 n=350K-500K (https://www.mersenneforum.org/showthread.php?t=11245)

 mdettweiler 2009-01-22 06:55

LLRnet IB8000 has completed 352K-355K; lresults emailed to Gary. :smile:

 gd_barnes 2009-01-23 10:06

Reserving n=400K-420K for port 8000. :smile:

 gd_barnes 2009-01-25 13:05

Reserving n=420K-430K for port 8000.

 gd_barnes 2009-01-27 09:13

Ian has processed the results for n=355K-360K to me. He will also be checking it against the sieve file and primes in several places. Once it is compared to the sieve file, it will be considered complete.

 gd_barnes 2009-01-27 09:41

As of 11 PM AZ time on Jan. 26th (6 AM GMT Jan. 27th), here is an updated count of all primes found in the 8th drive. It includes 2 primes not yet posted since they have not been submitted at top-5000 yet:

[code]
k primes
1400-1500 32
1500-1600 33
1600-1700 34

Total 99

k primes
1700-1800 18
1800-1900 24
1900-2000 18

Total 60
[/code]

Now, if you assume that the 23-3 start was just some random abberation and that it should be random from that point forward, subtracting that off still gives 76-57 in favor of k=1400-1700.

With the former k-range continuing to dominate, clearly k=1400-1700 must have a higher avg. weight than k=1700-2000. If anyone has time to look up and compute the avg. from [URL="http://www.rieselprime.org"]www.rieselprime.org[/URL], I'd be very curious to see it.

If it's not the weight, than something else is going on. If we are getting a higher percentage of primes from almost the same # of candidates in a sieve file for 2 distinct ranges, than we may have what we've been hoping for if this continues to run at such an alarming advantage for the smaller k-range...that is proof that these things may not be as random as we think they are! :smile:

Looking for this type of deviation from the norm is part of the reason the project was started...to have enough searched ranges without holes in them to prove a non-random deviation.

A little wishful thinking just yet but certainly worth further investigation.

Gary

 Mini-Geek 2009-01-27 12:49

[quote=gd_barnes;160632]With the former k-range continuing to dominate, clearly k=1400-1700 must have a higher avg. weight than k=1700-2000. If anyone has time to look up and compute the avg. from [URL="http://www.rieselprime.org"]www.rieselprime.org[/URL], I'd be very curious to see it.

If it's not the weight, than something else is going on. If we are getting a higher percentage of primes from almost the same # of candidates in a sieve file for 2 distinct ranges, than we may have what we've been hoping for if this continues to run at such an alarming advantage for the smaller k-range...that is proof that these things may not be as random as we think they are! :smile:

Looking for this type of deviation from the norm is part of the reason the project was started...to have enough searched ranges without holes in them to prove a non-random deviation.

A little wishful thinking just yet but certainly worth further investigation.

Gary[/quote]
The average weight of k=1400-1700 is 1805.600. For k=1700-2000 it is 1746.167. That's a difference of 59.433. I'm not terribly familiar with what weights mean what, but that seems like an insignificant difference to me. (For reference, my two k's I've reserved in the individual k drive are weighted 1463 and 1416, a difference of 47. Over 600K-1M, the difference in the number of candidates is only 137.) I think something else is going on, whether random or not.
Edit: By the way, k=1400-1700 has 3885 primes listed on that page, k=1700-2000 has 3789 primes. This correlates closely to the weights, not the recent bunching.

 gd_barnes 2009-01-28 15:49

I have now processed the results for this drive up to n=360K. I have compared primes found vs. the 1st post in this thread, the k=300-2000 page, and Karsten's 8th drive page. Everything looks good.

Karsten, the only inconsistency that I found was that you don't have k=1647 highlighted in blue on the k=300-2000 page. It appears that you are showing all k's where NPLB has found a new prime in blue. We found 1647*2^351262-1 prime.

Max or Ian, whenever you can, please process the results for n=360K-400K for this drive to me. The lowest k/n pair in the server is now n>400K.

Thanks,
Gary

 MyDogBuster 2009-01-28 18:25

[QUOTE]Max or Ian, whenever you can, please process the results for n=360K-400K for this drive to me. The lowest k/n pair in the server is now n>400K.[/QUOTE]

Max, I took care of it (I hope) Ian

Results emailed (6.5mb zipped) Whew

 mdettweiler 2009-01-28 22:42

[quote=MyDogBuster;160862]Max, I took care of it (I hope) Ian

Results emailed (6.5mb zipped) Whew[/quote]
Thanks--I'm glad that now there's at least three people who can process results now (myself, Gary, and you). And since I've been very busy lately and haven't had much time to devote to stuff like processing results, that is greatly appreciated. :smile:

 MyDogBuster 2009-01-28 22:49

[QUOTE]Thanks--I'm glad that now there's at least three people who can process results now (myself, Gary, and you).[/QUOTE]

Well, I'm still in training so what I did may be junk. LOL

 kar_bon 2009-01-29 09:41

[QUOTE=mdettweiler;160916]Thanks--I'm glad that now there's at least three people who can process results now (myself, Gary, and you). And since I've been very busy lately and haven't had much time to devote to stuff like processing results, that is greatly appreciated. :smile:[/QUOTE]

oh, you forgot the fourth one - ME!

but results for 7 (in words [b]S E V E N[/b]) servers it's hard to process them and update the pages too!

All times are UTC. The time now is 00:19.