20150827, 18:16  #1 
Nov 2008
3·167 Posts 
Primenet result not propagandizing to mersenne.ca
Hi,
Found another ECM factor for a 10m exponent yesterday (33 hours ago) M10858213 This doesn't appear in mersenne.ca which you can verify by clicking on the factor that links through. Ok, no problem, I'll just report it manually to mersenne.ca No, you can't, doesn't accept any results for any exponents less than 1000M. Which is a pain really as when it pulls over from priment it DOESN't pull over the sigma or number of curves.... 
20150827, 19:15  #2  
Serpentine Vermin Jar
Jul 2014
CCD_{16} Posts 
Quote:
I'm consulting with George on how to handle that one particular result... probably end up removing it entirely since at least that one is patently bogus. 

20150827, 20:57  #3  
Nov 2008
3×167 Posts 
Quote:


20150828, 03:42  #4 
Romulan Interpreter
Jun 2011
Thailand
21344_{8} Posts 
"propagandizing"? haha, who changed the title? 10 points! (hey, look mom, it is not underlined with red! we honestly had no idea this word is correct in English, in spite of the fact that we understand its meaning is not the same as propagating)
On topic: (Let's destroy James' holiday): There is a different problem here, I signaled few times in the past the fact that mersenne.ca is returning an error (factor unknown, or so), when one is looking for a factor which is not in DB. For example, if I look for q=23, and 23 (the factor) is not in DB, it will give me an error. It should be extremely easy to factor (q1)/2 and see if one of the prime factors of it can be the p, such as q divides 2^p1 (this is a single line in pari). And tell me "23 is a factor of M11" even if it never heared about 23 and 11. This because when some guy finds a factor, his first impulse is to click on that factor, to see how many bits it has. But then the link (automatically added by mersenne.org) will point to a mersenne.ca error page, because the factors are not yet synchronized (that can take from minutes to hours). I tried many times in the past to check for, or input, factors at/to mersenne.ca, but I gave up because it seemed too complicate to input the exponent too. I want to copy/paste the factor, that's it. Or click on it, on mersenne.org (i.e PrimeNet), as it is already working for "older" factors. Try it for a "new" factor (from the "cleared" report, for example), and mersenne.ca will crash. The site should be clever enough to find the exponent, and add the factor to its own database, enriching it (i.e. self learning) without waiting till that synchro process will occur, if the factor is valid, otherwise ignore it or tell me "you are idiot, this is not a valid factor". Of course, this can result in abuse, some idiot can send thousands of fake factors, keeping the site busy factorizing them. This can be timerestricted (no more than one q1 factorization every 102030 seconds, for the same IP address, for example; nobody finds new factors faster than that), or number_of_factorsrestricted, or have a capcha (yes, it is easier to input a capcha which you have it in front of you, than to go back to mersenne.org and look for the expo, look for "add factor" page on mersenne.ca, and copy/paste it there, then go back again and copy/paste the factor  this because the page where you look for factors is different than the page where you add factors). Last fiddled with by LaurV on 20150828 at 03:51 
20150828, 06:43  #5  
Serpentine Vermin Jar
Jul 2014
110011001101_{2} Posts 
Quote:
The funky "no factor" result was removed and James should be able to catch up with the mersenne.org data for the past 3 days now. I also reran the hourly stats just now to make sure "smok_bmv" went back to a more reasonable rank, not top 10 as the result of one bogus entry. Should be good. So if anyone noticed the hourly report for any of the rankings was updated at 6:36 UTC, that's why... it was me. 

20150829, 18:14  #6 
Nov 2008
3·167 Posts 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Odd result  swellman  YAFU  15  20160225 00:38 
I don't get the 'http://mersenne.org/primenet/'  joblack  Information & Answers  28  20100324 22:13 
Odd result  1997rj7  PrimeNet  2  20091204 08:48 
http://mersenne.org/primenet/ftops.txt messed up .  petrw1  PrimeNet  0  20070907 20:28 
GIMPS PrimeNet / www.mersenne.org server relocation  Old man PrimeNet  PrimeNet  25  20030423 20:24 