mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Information & Answers (https://www.mersenneforum.org/forumdisplay.php?f=38)
-   -   Am I doing this the right way? :) (https://www.mersenneforum.org/showthread.php?t=17933)

N871 2013-03-12 12:44

Am I doing this the right way? :)
 
Hello Everyone,

I have a short question.

I downloaded Prime95 to check 1 single number :)
I got however confused by how the software works, and suddenly
it start running automatically, checking M58xxxxxx.

I checked the worktodo.txt and changed this number in the number I
would like to check. Is this a correct way to search?

for the moment it has started again, and it says:

Mar 11 16:40] Starting trial factoring of Mxxxxxxx to 2^xx
[Mar 11 16:40] Trial factoring Mxxxxxxx to 2^xx.
[Mar 11 16:43] Trial factoring Mxxxxxxxxx to 2^xx is 0.67% complete. Time: 184.552 sec.


When will it start the real primality test?

thank you!

ET_ 2013-03-12 13:29

[QUOTE=N871;332954]Hello Everyone,

I have a short question.

I downloaded Prime95 to check 1 single number :)
I got however confused by how the software works, and suddenly
it start running automatically, checking M58xxxxxx.

I checked the worktodo.txt and changed this number in the number I
would like to check. Is this a correct way to search?

for the moment it has started again, and it says:

Mar 11 16:40] Starting trial factoring of Mxxxxxxx to 2^xx
[Mar 11 16:40] Trial factoring Mxxxxxxx to 2^xx.
[Mar 11 16:43] Trial factoring Mxxxxxxxxx to 2^xx is 0.67% complete. Time: 184.552 sec.


When will it start the real primality test?

thank you![/QUOTE]

The primality test will start after the trial-factoring test.

The trial-factoring test will test all possible divisors of Mxxxxxxxx between 2^(xx-1) and 2^xx, and will last 24/48 hours. This approach has been chosen to guarantee that no small factors divide your number, and the primality test is to be done.

If you get an exponent that you do not want to test, the best procedure is the following:

1 - Open your Prime95
2 - Go to the menu Test/Primenet
3 - Deselect the box "Use PrimeNet to get work and report results"
4 - Press OK
5 - Go to the menu Advanved/Unreserve Exponent...
6 - Write down the exponent you want to unreserve from PrimeNet and press OK

You may need to manual contact GIMPS (Advanced/Manual communication) to send your request

Luigi

N871 2013-03-12 13:35

Thanks a lot Luigi,

I hope it won't take months to test, my number is 10 digits long..
dual core 2500hz..

we'll see :)

So if I understand correctly I'm good to go. If I were to find a new Mersenne number, will
GIMPS be contacted automatically or not? just to be sure :)

thanks!

N871 2013-03-12 13:40

and a very last additional question:

Will Prime95 inform me immediately when he discovers that my number isn't a Mersennes number during the primality test? Or will it run for weeks to write it down in factors? Only to dissapoint me at the very end :)

firejuggler 2013-03-12 13:44

on an one modern core, it take from 3 week to one month to test a mersenne candidate from M58xxxxxx to M62xxxxxx.
10 digits? hmmm no, not month, years.... maybe decade

N871 2013-03-12 13:46

But once the primality test begins will I get a time estimation?

N871 2013-03-12 13:47

I think both my cores are doing the same thing...can I let them work together?

firejuggler 2013-03-12 13:49

if you are under windows, click test/status to have a rough estimate of the time it will take.

Mr. P-1 2013-03-12 14:08

[QUOTE=N871;332954]Mar 11 16:40] Starting trial factoring of Mxxxxxxx to 2^xx
[Mar 11 16:40] Trial factoring Mxxxxxxx to 2^xx.
[Mar 11 16:43] Trial factoring Mxxxxxxxxx to 2^xx is 0.67% complete. Time: 184.552 sec.[/QUOTE]

It's not clear how many digits your exponent has. If it is within the current testing range, then you may be testing an exponent which is assigned to someone else. This is called "poaching" and is strongly frowned on. You may also be testing an exponent which has already been tested.

Additionally, if all you did was change the exponent, and leave the other parts of the "Test=" line the same, then your assignment will be rejected the next time you talk to the server, because the assignment key (the hexadecimal string immediately following the equals sign) will be invalid. Delete this key. If the exponent is not otherwise assigned, you will be given another next time you talk to the server.

Also, you may be doing unnecessary factoring (or you may be omitting necessary factoring) because you will be using the parameters associated with the old exponent.

Go to [url=http://mersenne.org/report_exponent/]this page[/url] and enter your exponent. Check to see if it has P-1 done, and how far factored it is. Your worktodo line should look something like this:

Test=exponent,f,p

or

Test=key,exponent,f,p

where

key is the key given to you by the server for that exponent (not the one for the previous exponent.) If you are given such a key, don't lose it.

f is the highest number you see in a "no factor from ... to ..." line in your results query. There will be several such lines. You want the highest one.

p is 1 if there is a "P-1" line, else 0.

[quote]When will it start the real primality test?[/quote]

When it finishes trial and P-1 factoring, assuming these are necessary.

Alternatively, if P-1 is necessary, but you don't want to do it yourself. PM me your exponent, and I will do it for you. That way you can start the LL test sooner. If a factor turns up, I'll let you know and you can abandon the test. I can give about 12GB to the task, which may be more than you can.

Mr. P-1 2013-03-12 14:17

[QUOTE=N871;332958]I hope it won't take months to test, my number is 10 digits long..
dual core 2500hz..[/QUOTE]

I'd not seen this when I wrote my previous reply.

I don't know if prime95 can even handle numbers this size. Even if it can, it would take decades to complete the test, and your chances of doing so without error are tiny. Bear in mind that a single uncorrected error at any point would invalidate the result of the whole computation.

No test of a number this size has ever been completed. There is a reason for this.

Needless to say my offer to do the P-1 doesn't apply to billion digit Mersennes.

ET_ 2013-03-12 15:01

Just to be clear:

is your number or your exponent that has 10 digits?
Is your number a Mersenne number (like 2^p-1)?

Luigi


All times are UTC. The time now is 22:50.

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