Other

drummers-lowrise

Message boards : Project Staging Area : Prime/PRP in GFN results

 Subscribe SortOldest firstNewest firstHighest rated posts first
Author Message
stream
Volunteer moderator
Volunteer developer
Volunteer tester

Joined: 1 Mar 14
Posts: 1004
ID: 301928
Credit: 543,187,198
RAC: 7

Message 85843 - Posted: 30 May 2015 | 20:20:18 UTC

May be this question is stupid, but in GFN "Primes by user" tables I see a column "Primes/PRP" and titles like "User xxxx has found N primes and K PRPs". What does PRP mean is this context? I know it usually stands for "probable prime". I'm curious because there is a one such entry in GFN32768 results:

6656034^32768+1 PRP

but when I run it through latest LLR, it said that number is true prime:

stream@ubuntuserver:~/boincdata/projects/www.primegrid.com\$ ./llr64 -d -q"6656034^32768+1" Base factorized as : 2*3*7*11*14407 Base prime factor(s) taken : 14407 Starting N-1 prime test of 6656034^32768+1 Using generic reduction FMA3 FFT length 72K, Pass1=384, Pass2=192, a = 3 6656034^32768+1 may be prime, trying to compute gcd's 3^((N-1)/14407)-1 is coprime to N! 6656034^32768+1 is prime! (223584 decimal digits) Time : 1490.407 sec.

Am I missing something or it's just a wrong entry in the table?

Michael Goetz
Volunteer moderator

Joined: 21 Jan 10
Posts: 13827
ID: 53948
Credit: 374,315,866
RAC: 400,632

Message 85844 - Posted: 30 May 2015 | 22:09:37 UTC

PRP stands for probable prime.

For all intents and purposes, there's no difference in this context and they should be considered to be the same thing. The odds of a Genefer finding a PRP that is actually composite are minute. So effectively, there's no difference.

The PRPNet software makes a distinction between the user finding a PRP (using Genefer) vs the user finding a prime (using Genefer to find the PRP followed by PFGW proving the primality or by just using PFGW for the entire test), but we prefer that users just use Genefer and we'll verify the prime on our end. It's much faster that way and eliminates the risk of something going wrong with the PFGW test. PFGW can't checkpoint when doing that kind of test, so it's risky and we'd rather people didn't run it.

(Note that it's possible that more recent versions of PRPNet use LLR instead of PFGW, but we still prefer that users just run the Genefer test with PFGW and LLR commented out. There's less to go wrong.)
____________
My lucky number is 75898524288+1

stream
Volunteer moderator
Volunteer developer
Volunteer tester

Joined: 1 Mar 14
Posts: 1004
ID: 301928
Credit: 543,187,198
RAC: 7

Message 85850 - Posted: 31 May 2015 | 18:49:05 UTC - in response to Message 85844.

Thanks, Michael. There are total 3 of such results, marked as "PRP"

6656034^32768+1
3503138^65536+1
773620^262144+1

and all of them are true primes (first one is shown above, other two can be found in top 5000). Since ALL other GFN results were marked as "Prime", it looked suspicious at first glance, like a glitch in database or a field forgotten to be set manually. Thanks again for explanation.

Message boards : Project Staging Area : Prime/PRP in GFN results