mersenneforum.org  

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

Closed Thread
 
Thread Tools
Old 2005-12-17, 20:45   #67
akruppa
 
akruppa's Avatar
 
"Nancy"
Aug 2002
Alexandria

25·7·11 Posts
Default

Quote:
Originally Posted by pacionet
If I understand well, in the past there were some false positive ?
They were due to software or hardware problems ?
There was once. The exact reason was never found out, but George discovered that if one particular variable (iirc, a DWT weight of type double) was zeroed out during execution, all the computed residues during the LL test could become zero - so the last residue would erroneously cause a prime to be reported. The code has been hardened against such behaviour since, and the usual verification of the last 2000 iterations from a save file makes this kind of error impossible to go undetected.

Alex
akruppa is offline  
Old 2005-12-17, 20:56   #68
tha
 
tha's Avatar
 
Dec 2002

31316 Posts
Default

Quote:
Originally Posted by ewmayer
I just got a PM from George - he has no e-mail access but is apparently able to login to his merseneforum account, although I can't vouch for the speed of his internet connection.
Note to self: whenever starting up something succesful and gathering followers remember to make a list beforehand of places to hide occasionally spend holidays that have absolute no TCP/IP access.
tha is offline  
Old 2005-12-17, 21:22   #69
moo
 
moo's Avatar
 
Jul 2004
Nowhere

80910 Posts
Default

not to self just leave all electronics at home includeing the urge to use computers....

P.S. That company needs to start getting ready to print another poster...

Last fiddled with by moo on 2005-12-17 at 21:23
moo is offline  
Old 2005-12-17, 21:48   #70
TTn
 

132118 Posts
Default 0's in the residue

Quote:
There was once. The exact reason was never found out, but George discovered that if one particular variable (iirc, a DWT weight of type double) was zeroed out during execution, all the computed residues during the LL test could become zero - so the last residue would erroneously cause a prime to be reported. The code has been hardened against such behaviour since, and the usual verification of the last 2000 iterations from a save file makes this kind of error impossible to go undetected.
Alex
Well spoken.

There also have been numerous problems with 16bit-64bit conversion.
No one officially has got it down right.
REFERENCES

a. Tech Ed conference 2002
b. VS.NET documentation and videos with in depth explanations.
c. History channel, also showed that N.A.S.A, and the European space administration could not figure out why there were so many zeroes in the residue.

Bam!

Let me guess I banned again, for talking the truth.
Shane F.

Last fiddled with by TTn on 2005-12-17 at 21:50
 
Old 2005-12-17, 22:16   #71
flava
 
flava's Avatar
 
Feb 2003

2·59 Posts
Default

Looks like someone has an agenda
flava is offline  
Old 2005-12-17, 22:19   #72
TTn
 

29×113 Posts
Default

Important note:

I forgot about the theory proposed (c.), that the old software, could not handle the new hardware, because of the speed increases.
This was ofcourse an unforseen thing.

The theory(IMHO) makes moderate sense, as Windows 3.1 (roots of p95) was based on synchronus behaviour, with very few API, while 32+ is geared to asynchronus behaviour(multi-tasking).
Then I'd have to go on, to tell you I told you so.... "new syntax"
 
Old 2005-12-18, 02:16   #73
lycorn
 
lycorn's Avatar
 
Sep 2002
Oeiras, Portugal

58316 Posts
Default

If my guess is correct:

1. it won´t be a 10M digit prime
2. the discoverer really deserves to have found a prime...

What you guys think about it?
lycorn is offline  
Old 2005-12-18, 03:44   #74
Fusion_power
 
Fusion_power's Avatar
 
Aug 2003
Snicker, AL

23×3×52 Posts
Default

It took less than an hour to figure out what the correct candidate is. I won't post it and neither should anyone else until a full verification has been done. Be patient.

As for finding the next Mp, a little bit of sleuthing will tell you that gimps has more computers working but each individual test takes longer as the numbers get larger. Each time a prime is found, a significant proportion of new computers are added. The two effects tend to cancel each other.

Fusion
Fusion_power is offline  
Old 2005-12-18, 08:20   #75
Citrix
 
Citrix's Avatar
 
Jun 2003

32·52·7 Posts
Default

Quote:
Originally Posted by Fusion_power
It took less than an hour to figure out what the correct candidate is. I won't post it and neither should anyone else until a full verification has been done. Be patient.

As for finding the next Mp, a little bit of sleuthing will tell you that gimps has more computers working but each individual test takes longer as the numbers get larger. Each time a prime is found, a significant proportion of new computers are added. The two effects tend to cancel each other.

Fusion
Congrats to George and the GIMPS project.

Is it above 10 M digits?

Citrix
Citrix is offline  
Old 2005-12-18, 08:27   #76
Numbers
 
Numbers's Avatar
 
Jun 2005
Near Beetlegeuse

22×97 Posts
Default

I am obviously very naïve because I’m afraid I do not understand the need for secrecy. My current exponent is due to finish in a few days and if Prime95 told me it was a probable prime I would be straight onto the forum to say so. Can anyone give a compelling reason why I should not?
Numbers is offline  
Old 2005-12-18, 08:33   #77
akruppa
 
akruppa's Avatar
 
"Nancy"
Aug 2002
Alexandria

25×7×11 Posts
Default

We can't be really sure it is an actual prime until an independent verification run is completed. It is possible, although very unlikely, that a false positive is reported in the first run. We used to think that it even was practically impossible - until it happened. The secrecy saved GIMPS a lot of embarassment back then. If the exponent had been blurted out before verification, the discovery would have received a lot of press only to be followed up with a report of retraction of the announcement. Double-checking results before publishing them is simply a matter of scientific prudence.

Alex
akruppa is offline  
Closed Thread

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Are Bitcoins Prime Related a1call Miscellaneous Math 23 2020-09-17 13:17
Holy new Mersenne prime, Batman! (M47 related) ixfd64 News 901 2008-11-10 19:29
Another new prime (M42-Related) Uncwilly News 132 2005-05-10 19:47
Mersenne prime related shirts and other items adpowers Lounge 40 2004-08-12 22:05
some prime-related trick questions ixfd64 Puzzles 2 2003-09-23 12:53

All times are UTC. The time now is 01:09.

Sun Nov 1 01:09:38 UTC 2020 up 51 days, 22:20, 3 users, load averages: 1.31, 1.59, 1.68

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, 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.