mersenneforum.org  

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

Reply
 
Thread Tools
Old 2009-08-12, 17:29   #1
S485122
 
S485122's Avatar
 
Sep 2006
Brussels, Belgium

2·5·167 Posts
Default same exponent reassigned for triple-check

Just after completing a double-check I was reassigned the same exponent (my returned result differed from the result previously returned by another user.) I unreserved the exponent and it was reassigned to me once more. I escaped the problem by getting more assignments by increasing the number of “Days of work to queue up” before unreserving the litigious exponent. See the excerpt of the log file in attachment.

It seems the check to insure an exponent is not reassigned to the same user for double-checking (triple-checking in fact) failed.

Jacob
Attached Files
File Type: txt prime.log.txt (2.9 KB, 146 views)
S485122 is offline   Reply With Quote
Old 2009-08-13, 03:32   #2
lfm
 
lfm's Avatar
 
Jul 2006
Calgary

1101010012 Posts
Default

Quote:
Originally Posted by S485122 View Post
It seems the check to insure an exponent is not reassigned to the same user for double-checking (triple-checking in fact) failed.
There is, as you discovered, no such check, and I don't think there ever has been. The double checks include a random "offset" which allows different double checks to be run on the same system.
lfm is offline   Reply With Quote
Old 2009-08-13, 10:06   #3
Brian-E
 
Brian-E's Avatar
 
"Brian"
Jul 2007
The Netherlands

7·467 Posts
Default

Quote:
Originally Posted by lfm View Post
The double checks include a random "offset" which allows different double checks to be run on the same system.
Yes but if Jacob agreed to test that exponent for the third time and the third result agreed with his own previous test, then the exponent would still be considered insufficiently tested because it would not have two agreeing tests by two independent testers which is what GIMPS requires. The random offset at the start is meant to eliminate possible software errors which could allow two different testers to turn in the same incorrect result; it is not intended to cancel the need for independent verification.

Last fiddled with by Brian-E on 2009-08-13 at 10:11 Reason: clarified last sentence
Brian-E is offline   Reply With Quote
Old 2009-08-13, 21:17   #4
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

10111111111112 Posts
Default

Quote:
Originally Posted by Brian-E View Post
The random offset at the start is meant to eliminate possible software errors which could allow two different testers to turn in the same incorrect result; it is not intended to cancel the need for independent verification.
Maybe not?

http://mersenne.org/report_LL/?exp_l...B1=Get+LL+data

"curtisc" did both the fist and second tests. Will that exponent be assigned for third check also?
retina is online now   Reply With Quote
Old 2009-08-14, 00:29   #5
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

22·3·641 Posts
Default

Quote:
Originally Posted by retina View Post
Maybe not?
There's intention ... and then there's implementation.
cheesehead is offline   Reply With Quote
Old 2009-08-14, 02:43   #6
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

6,143 Posts
Default

Quote:
Originally Posted by cheesehead View Post
There's intention ... and then there's implementation.
Touché.
retina is online now   Reply With Quote
Old 2009-08-14, 20:30   #7
sichase
 
sichase's Avatar
 
Dec 2008
Sunny Northern California

718 Posts
Default

curtisc did them both, but on two different pieces of hardware. I think that a triple-check will only be assigned if the same hardware is used twice.
sichase is offline   Reply With Quote
Old 2009-08-14, 21:51   #8
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

614310 Posts
Default

Quote:
Originally Posted by sichase View Post
curtisc did them both, but on two different pieces of hardware. I think that a triple-check will only be assigned if the same hardware is used twice.
Yes, I am sure we can trust curtisc, but that is not the point. Another less trustworthy user might fake the second result (for whatever reason) and no one would know.
retina is online now   Reply With Quote
Old 2009-08-16, 23:16   #9
sichase
 
sichase's Avatar
 
Dec 2008
Sunny Northern California

3·19 Posts
Default

Quote:
Originally Posted by retina View Post
Yes, I am sure we can trust curtisc, but that is not the point. Another less trustworthy user might fake the second result (for whatever reason) and no one would know.
It would be easy enough for a malicious user just to create a second user name to beat such a naive safeguard. I don't know the details, but I assume that the protection provided by the second test is robust against such trivial gaming.
sichase is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
error during Jacobi check on 330,000,000+ exponent evanh Hardware 5 2018-02-20 03:46
Triple check required houding PrimeNet 14 2015-12-21 09:34
mprime ROUND OFF ERROR: Triple-check advised? Bdot Software 5 2012-12-22 22:34
An unnecessary triple check ? S485122 PrimeNet 9 2008-09-24 13:41
Two priority triple-check exponents GP2 Completed Missions 13 2004-08-12 00:41

All times are UTC. The time now is 06:19.

Sun May 16 06:19:24 UTC 2021 up 38 days, 1 hr, 0 users, load averages: 2.60, 2.34, 2.18

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.