mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Data

Reply
 
Thread Tools
Old 2003-10-02, 00:08   #12
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

11100101111112 Posts
Default

22851593 is way beyond the range of exponents PrimeNet is handing out for LL testing. The only way I know of that this exponent could get in a worktodo.ini file would be for a user to manually add it.
Prime95 is online now   Reply With Quote
Old 2003-10-02, 05:41   #13
GP2
 
GP2's Avatar
 
Sep 2003

32·7·41 Posts
Default

Quote:
Originally posted by Prime95
22851593 is way beyond the range of exponents PrimeNet is handing out for LL testing. The only way I know of that this exponent could get in a worktodo.ini file would be for a user to manually add it.
I emailed Team Siegert and the response was:


No, I did not do that, and I don't understand why ECE1 would have been LL-testing that exponent.



Siegert does have a few factoring assignments in the 22M range, but that particular exponent didn't have an F in status.txt (and in any case factoring wouldn't have made any more sense since it was already factored back in August, and this latest exponent had an "assigned" date of Sept 30).

It's a mystery.

Last fiddled with by GP2 on 2003-10-02 at 05:42
GP2 is offline   Reply With Quote
Old 2003-10-06, 13:53   #14
GP2
 
GP2's Avatar
 
Sep 2003

32·7·41 Posts
Default

One more new one (I already resubmitted it to Primenet):


Exponents assigned (LL testing) that already have a factor?

-------
factors
-------
21876539,1094145315290536903

-------
STATUS_L.TXT
-------
21876539, ,60,,0.8,58.2,87.2,05-Oct-03 02:41,05-Oct-03 02:37,Team_Prime_Rib,the_one


In an old cleared.txt file this appeared as:

21876539,60, F,1094145315290536903,30-Aug-03 20:28,wabbit,factoring


Any ideas how this one got assigned? It's also beyond the leading edge of first-time checking.
GP2 is offline   Reply With Quote
Old 2003-10-07, 00:17   #15
geoff
 
geoff's Avatar
 
Mar 2003
New Zealand

13·89 Posts
Default

Quote:
In an old cleared.txt file this appeared as:


21876539,60, F,1094145315290536903,30-Aug-03 20:28,wabbit,factoring


Any ideas how this one got assigned? It's also beyond the leading edge of first-time checking.

From the look of the lines in cleared text it seems wabbit has some machines with Factoroverride=62 or similar in their prime.ini, as there are hundreds of factors being found but hardly any greater than 2^62.

I read somewhere else that using Factoroverride with primenet assignments causes problems, so this could be related.
geoff is offline   Reply With Quote
Old 2003-10-18, 18:31   #16
GP2
 
GP2's Avatar
 
Sep 2003

32·7·41 Posts
Default

More weirdness:

Code:
Exponents assigned (LL testing) that already have a factor?
 
-------
factors
-------
10300013,535658912871414046645321
33459763,16228318390260061537
 
 
-------
STATUS_L.TXT
-------
10300013,D ,64,,4.3,16.7,76.7,,14-Oct-03 08:31,.,C81F6D02B
33459763,  ,64,1,3.8,60.2,84.2,,14-Oct-03 20:35,S101148,C4B90EFBA

The factor for the 33M exponent was discovered July 13.

The factor for the 10M exponent first appears in the Sept 29 version of the data files, but I can't find any record of it appearing in a cleared.txt file.

I wonder how these got assigned recently (Oct 14).

The factors have already been re-reported via the manual testing form.


There's no other record of user S101148 or computer-id C4B90EFBA anywhere else: not in HRF5.TXT (list of all users), not in any entry in LUCAS_V.TXT, BAD or HRF3.TXT, not in any other line in status.txt or cleared.txt
GP2 is offline   Reply With Quote
Old 2003-10-19, 05:51   #17
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

11110000011002 Posts
Default

Quote:
Originally posted by geoff
I read somewhere else that using Factoroverride with primenet assignments causes problems, so this could be related.
AFAIK, the only problem associated with using FactorOverride with PrimeNet assignments is that the FactorOverride gets ignored! It doesn't corrupt assignments or cause the type of problems being discussed here.

From a previous posting of mine:

If I interpret the source code correctly, Prime95 module commonc.c ignores FactorOverride _if_ the user has selected the "Use PrimeNet to get work and report results" box checked in the "Configure PrimeNet" options.

AFAIK this is not because of any technical limitation, but because George wanted PrimeNet assignments to use the default limits for trial factoring (which I agree is a good idea). My guess is that his undoc.txt statement "This feature should not be used with the Primenet server" may have been left overly broad in order to discourage unknowledgable users from messing with default limits without good reason.

The results line for trial factoring doesn't say anything about whether or not FactorOverride has been used, and the default trial factoring limits have sometimes been changed between past versions of Prime95, so I don't think PrimeNet itself really cares or even knows.
cheesehead is offline   Reply With Quote
Old 2003-10-19, 08:09   #18
geoff
 
geoff's Avatar
 
Mar 2003
New Zealand

100100001012 Posts
Default

Quote:
AFAIK, the only problem associated with using FactorOverride with PrimeNet assignments is that the FactorOverride gets ignored! It doesn't corrupt assignments or cause the type of problems being discussed here.
What does the server do if a factoring assignment is returned without finding a factor but not factored to the normal bit depth, does it hand it out again as a factoring assignment?
geoff is offline   Reply With Quote
Old 2003-10-19, 12:07   #19
Complex33
 
Complex33's Avatar
 
Aug 2002
Texas

5×31 Posts
Default

From what I can tell is if a factoring assignment is not completed to the default depth, the exponent is treated as though no new factoring has been done and is re-released at its previous bit depth. This is why I have concern about the factoring that wabbit has been doing on the upper TF exponents on primenet. It seems as though he has invested a great deal in factoring those numbers to 62 bits in order to find fast factors for credit but then primenet has to re-release these exponents and the next user is repeating work up to 62 when no factor will be found. Is this right?
Complex33 is offline   Reply With Quote
Old 2003-10-19, 14:44   #20
GP2
 
GP2's Avatar
 
Sep 2003

32·7·41 Posts
Default

Quote:
Originally posted by Complex33
It seems as though he has invested a great deal in factoring those numbers to 62 bits in order to find fast factors for credit but then primenet has to re-release these exponents and the next user is repeating work up to 62 when no factor will be found. Is this right?
When Primenet assigns you an exponent, it tells you how far it's been trial-factored already (second parameter in the Test= or DoubleCheck= line).

And the posts in the Lone Mersenne Hunters forum consist mostly of "I'm factoring to 2^n for range xM-yM". So this is an organized and approved activity, no?
GP2 is offline   Reply With Quote
Old 2003-10-19, 18:15   #21
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

19·499 Posts
Default Factoring...

Quote:
Originally posted by GP2
And the posts in the Lone Mersenne Hunters forum consist mostly of "I'm factoring to 2^n for range xM-yM". So this is an organized and approved activity, no?
Not quite, actually...

I'm working in the high-end of ranges which are in Primenet, which the Lone Hunters don't do. I'm not reporting back to Primenet any results except factors found, to be sure I don't cause problems for the server.

However, from time to time I send my results to George (via CSV flat files) to import into his dataset so that others are not assigned work which is redundant.
chalsall is online now   Reply With Quote
Old 2003-10-21, 03:58   #22
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

22·3·641 Posts
Default

Quote:
Originally posted by geoff
What does the server do if a factoring assignment is returned without finding a factor but not factored to the normal bit depth, does it hand it out again as a factoring assignment?
Good point. After someone has received factoring assignments from PrimeNet, she/he can turn off "Use PrimeNet to get work and report results" and then the "FactorOverride=" will not be ignored.

Quote:
Originally posted by Complex33
From what I can tell is if a factoring assignment is not completed to the default depth, the exponent is treated as though no new factoring has been done and is re-released at its previous bit depth.
... which is a failure of PrimeNet to properly handle the returned result. (This situation could occur if the default bit limit were changed between time of assignment and return of the result ... not that that is likely to happen again soon.)

Quote:
This is why I have concern about the factoring that wabbit has been doing on the upper TF exponents on primenet. It seems as though he has invested a great deal in factoring those numbers to 62 bits in order to find fast factors for credit but then primenet has to re-release these exponents and the next user is repeating work up to 62 when no factor will be found. Is this right?
Apparently so, as long as PrimeNet fails to properly update the bit level.

Quote:
Originally posted by GP2
And the posts in the Lone Mersenne Hunters forum consist mostly of "I'm factoring to 2^n for range xM-yM". So this is an organized and approved activity, no?
But the LMH are not working in ranges that PrimeNet is assigning.
cheesehead is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Composite P-1 factors not showing up under recently cleared exponents? ixfd64 PrimeNet 2 2018-02-28 07:54
[script] v4-style table of cleared exponents ixfd64 PrimeNet 1 2010-06-14 16:24
We lost some Cleared Exponents??? petrw1 PrimeNet 1 2007-04-30 17:35
Cleared exponents that never made it into data files GP2 Data 14 2003-09-16 03:07
Anomalous entries in assigned & cleared files GP2 Data 2 2003-09-09 14:40

All times are UTC. The time now is 00:26.

Sun Mar 7 00:26:57 UTC 2021 up 93 days, 20:38, 0 users, load averages: 1.93, 1.73, 1.77

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

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.