mersenneforum.org  

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

Closed Thread
 
Thread Tools
Old 2004-08-11, 21:11   #12
PrimeCruncher
 
PrimeCruncher's Avatar
 
Sep 2003
Borg HQ, Delta Quadrant

2·33·13 Posts
Default

Quote:
Originally Posted by JFB
I don't understand, PrimeCruncher. Didn't you say you haven't tested it ?
Mmm... Maybe I'm too newB for that matter...
I will think about it some time.
Anyway thanks to both of you !
JFB.
I haven't tested it going BEYOND the standard depth, which is what I thought you were asking.
PrimeCruncher is offline  
Old 2004-08-19, 11:26   #13
JFB
 
Jun 2004
France [UTC+2]

2×7 Posts
Default

Well, as Axel Fox said it, FactorOverride parameter is ignored when you use Primenet. I tested it with Prime95 23.8.1 on windows XP pro.

Thus my question is still pending :
does anybody know how to factorize deeper than default when running Prime95 primality test AND using Primenet ?
JFB is offline  
Old 2004-08-19, 11:49   #14
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

7268 Posts
Default

If you really want to factor an exponent higher, you could use the "advanced factor" option.

I used it to factor an exponent higher than what prime95 wanted to factor it. (Refer to "Server gave me same exponent twice" in the server forum).

However, you will still have to manually edit the worktodo.ini file, however at least that way you can still have "use primenet" checked.
dave_0273 is offline  
Old 2004-08-19, 12:12   #15
JFB
 
Jun 2004
France [UTC+2]

2·7 Posts
Default

With Prime95, is it possible to do the sequence :

- have a LLtest assigned by Primenet (Test=... in worktodo.ini)
- factorize deeper than default
- if no factor continue with standard P-1 factoring and LL primality test
- return result to Primenet

?

JFB
JFB is offline  
Old 2004-08-19, 17:11   #16
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

2×5×47 Posts
Default

Yes. However, you will still have to manually edit the worktodo file in order to do it.

1) Recieve your exponent from primenet
2) Now that you have your exponent, stop and exit prime95.
3) Insert the "Advancedfactor=" line at the TOP of your worktodo file
4) Start up prime95 again

(Now this is the part that you will need to experiment with. I am 90% sure that this will work, but just give it a test run and see what happens.)

5) If the program finds a factor in the "advancedfactor" stage, prime95 should automatically contact the server to tell it about the factor. At that time, prime95 should kick the LL test line out of your worktodo file seeing as that exponent has already got a factor.

6) The server should then assign a new exponent to LL test.
7) When you check back on prime95 later on, you will see that it is now working on a new exponent. Even though the LL test is a couple of thousand interations in (or still doing the p-1 test) you can still stop prime95 and insert the advancedfactor at the top of worktodo. Then prime95 will stop the LL test, factor it first, and then continue the LL test from where it left off earlier.

On the other hand, if a factor was not found in step 5, the program will report that no factor was found and then continue on with the LL test.

I hope that this helps

Last fiddled with by dave_0273 on 2004-08-19 at 17:11
dave_0273 is offline  
Old 2004-08-20, 07:49   #17
JFB
 
Jun 2004
France [UTC+2]

11102 Posts
Default

Thanks Dave, it helps.
But ...
1/ the Advancedfactor feature is "likely to be deleted in a futur release" (in undoc.txt)
2/ let's consider this sequence :
- worktodo.ini : Advancedfactor ... ==> done, assume it found no factor
- worktodo.ini : Test ... ==> ? do the trial factoring twice ?

I feel like your avatar ...
JFB
JFB is offline  
Old 2004-08-20, 10:31   #18
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

2·5·47 Posts
Default

Quote:
Originally Posted by JFB
2/ let's consider this sequence :
- worktodo.ini : Advancedfactor ... ==> done, assume it found no factor
- worktodo.ini : Test ... ==> ? do the trial factoring twice ?
When an exponent is assigned to you as an LL test, it has already been trial factored up to its required depth. Therefore after it has completed the "advancedfactor" step, that is the end of trial factoring. When it moves onto the test line, it will start the p-1 test, and then if no factor is found will move straight onto the LL test.

In short...
If you are getting non 10mil exponents from the server as first time LL tests or double checks, there will be no "trial factoring" done in the "test=" line. That would have already been completed.

/edit starts here

I just re-read your first post and realised that you are working on 10mil exponents. Therefore, you will have to add another line to prime.ini. From undoc.txt

Code:
You can force prime95 to skip the trial factoring step prior to
running a Lucas-Lehmer test.  In prime.ini add this line:
	SkipTrialFactoring=1
Therefore, when it gets to the "Test=" line, it will not do any more trial factoring on the exponent.

Last fiddled with by dave_0273 on 2004-08-20 at 10:35
dave_0273 is offline  
Old 2004-08-21, 13:12   #19
JFB
 
Jun 2004
France [UTC+2]

2·7 Posts
Default

Thanks again Dave.

I feel that both solutions (FactorOverride & AdvancedFactor)
are not compatible with the design of Primenet.
Therefore it's unlikely that I use them.

However, I still wonder how those entries in Primenet cleared.txt are produced :
- successfull higher-than-default factorizations for 10M digits in August 2004:
Code:
          34106339  95   F    44498800115800743380399887577  20-Aug-04 07:43
          34072667  82   F        6788467951496287519001311  20-Aug-04 08:37
         34109419  71   F           2585513676957889796383  20-Aug-04 05:48
         34078691  74   F          16798666949116291067161  19-Aug-04 04:55
         34108619  77   F         202392820834359951199769  18-Aug-04 07:56
        34104223  70   F           1582809736990555899193  17-Aug-04 06:01
        34102787  93   F     9098482339355640635118005113  16-Aug-04 16:36
        34031143  72   F           5731604239721652326497  14-Aug-04 04:15
        34054343  70   F           1656849570126477484489  14-Aug-04 10:01
        34099423  97   F   213474212032872114112269659543  13-Aug-04 01:37
        34103591  84   F       26516580398164700900130497  13-Aug-04 17:04
        34099789  72   F           4448786183354641912337  13-Aug-04 10:22
        34092827  82   F        5409856357736134619194687  12-Aug-04 11:45
        34096607  78   F         235447854405595499929081  12-Aug-04 23:21
        34094629  76   F          91054747635313368077113  12-Aug-04 13:16
        34098569  74   F          17920473480671643148999  12-Aug-04 17:04
        34095617  86   F      108464588320835914323557239  09-Aug-04 13:25
        34088881 103   F  9650611843244071939107824508149  08-Aug-04 07:57
        34096897  75   F          27890759839597600749313  07-Aug-04 23:14
        34086889  72   F           5014251748474414282999  06-Aug-04 18:07
        34056377  73   F           9902750430318580875023  03-Aug-04 14:22
        34084487  82   F        6653668203102342242587537  02-Aug-04 03:27
        34076893  73   F          12348528709298673483169  02-Aug-04 08:24
        34082497  71   F           2075415882871271986447  01-Aug-04 18:53
- ?? (unsuccessful ?)
Code:
 35062597 73	 0x3A8C5067AB26B8__			 26-Dec-03 12:27
 33302131 70	 0x19BE010D1785C3__			 25-Jun-04 15:44
 33306257 70	 0x44BF910A3F558C__			 07-Apr-04 18:19
 33329081 70	 0xC010CC57795A86__			 09-Mar-04 10:44
 33520813 70	 0x315337083031E0__			 07-Jun-04 02:51
 33637147 70	 0x141201C7FF3532__			 03-Jun-04 14:46
 33810761 70	 0x55CBE0D5446F68__			 21-Apr-04 10:56
To any owner of one of these exponents : explanations are welcome, please.

JFB
JFB is offline  
Old 2004-08-21, 13:15   #20
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

25·13·23 Posts
Default

Are you sure that those are not P-1 factors? because to take a number to 90 would take longer than the L-L (I believe).
Uncwilly is offline  
Old 2004-08-21, 13:28   #21
dave_0273
 
dave_0273's Avatar
 
Oct 2003
Australia, Brisbane

7268 Posts
Default

All those exponents that you see have higher bit factors (<70 bits) were all found through p-1 factoring.

Quote:
Originally Posted by JFB
- ?? (unsuccessful ?)

Code:
35062597 73 0x3A8C5067AB26B8__ 26-Dec-03 12:27
33302131 70 0x19BE010D1785C3__ 25-Jun-04 15:44
33306257 70 0x44BF910A3F558C__ 07-Apr-04 18:19
33329081 70 0xC010CC57795A86__ 09-Mar-04 10:44
33520813 70 0x315337083031E0__ 07-Jun-04 02:51
33637147 70 0x141201C7FF3532__ 03-Jun-04 14:46
33810761 70 0x55CBE0D5446F68__ 21-Apr-04 10:56
These are the residues of of the LL test. Unless the residue is 0, the mersenne is NOT prime. We keep the residue because latter on a double check will be done to make sure that there was no errors in the test.

Although we now know that these are not primes, we still don't know any of their factors.
dave_0273 is offline  
Old 2004-08-21, 13:58   #22
JFB
 
Jun 2004
France [UTC+2]

2×7 Posts
Default

OK I understand, thanks to both of you !
I've got one more question :
does the depth of P-1 factoring depend on the amount of ram it gets ?
For instance, 34,088,881 reached 103 F, sounds huge...

JF
JFB is offline  
Closed Thread

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
What is Trial Factoring? Unregistered Information & Answers 5 2012-08-02 03:47
How much Trial Factoring to do? odin Software 4 2010-08-08 20:23
How far to do trial factoring S485122 PrimeNet 1 2007-09-06 00:52
How to only do Trial Factoring? michael Software 23 2004-01-06 08:54
About trial factoring gbvalor Math 4 2003-05-22 02:04

All times are UTC. The time now is 10:17.

Thu May 13 10:17:57 UTC 2021 up 35 days, 4:58, 1 user, load averages: 1.93, 3.22, 2.87

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.