Malformed PRP-CF assignment from server

ric · 2021-05-09, 10:20

Got assigned

Code:

PRP=<aid>,1,2,3215747,-1,99,0,3,1,"4457025343,185822885311153245017,185822885311153245017"

from Primenet, and mprime complained with

Code:

185822885311153245017 does not divide exponent

which is clearly not true, and discarded the worktodo line.

Double-checked the assignment line from my assignments list, and realized that in there the recent factor was contained twice (hence the warning); wandering whether this depends from it being reported twice to the server. Removed the duplicate from the assignment line, changed the '99,0,3,1' to '99,0,5,1', now happily crunching.

I'm pretty confident this is a very corner-case, so I'm leaving this note mainly as a self-reminder.

ATH · 2021-05-09, 11:00

Yes, this is a known error with unknown cause:
https://mersenneforum.org/showpost.p...&postcount=258
https://mersenneforum.org/showpost.p...&postcount=259

ewmayer · 2021-05-20, 20:21

Quote:

Originally Posted by ric

Removed the duplicate from the assignment line, changed the '99,0,3,1' to '99,0,5,1', now happily crunching.

PRP-CF n00b here - why change the PRP base from 3 to 5?

ATH · 2021-05-20, 21:59

PRP-CF are usually run with type 5, because then if a new factor is found, only 1 more test is needed, because the residue will be the same with the new factor.

https://mersenneforum.org/showthread.php?t=25841

ewmayer · 2021-05-20, 22:13

Quote:

Originally Posted by ATH

PRP-CF are usually run with type 5, because then if a new factor is found, only 1 more test is needed, because the residue will be the same with the new factor.

https://mersenneforum.org/showthread.php?t=25841

Yes, but in e.g. PRP=<aid>,1,2,3215747,-1,99,0,3,1..., 3 is the PRP base and 1 is the PRP-test (and cofactor-post-test) residue type, no? The OP said he changed the 3 to 5, which would give a base-5 PRP-test with cofactor-test residue type 1, unless I need my eyes examined (which is certainly possible).

kruoli · 2021-05-21, 08:34

Yes, the test does show up as using base 5 and type 1 on the results page: 3215747.

ATH · 2021-05-21, 15:20

Quote:

Originally Posted by ewmayer

Yes, but in e.g. PRP=<aid>,1,2,3215747,-1,99,0,3,1..., 3 is the PRP base and 1 is the PRP-test (and cofactor-post-test) residue type, no? The OP said he changed the 3 to 5, which would give a base-5 PRP-test with cofactor-test residue type 1, unless I need my eyes examined (which is certainly possible).

Sorry, I did not notice it was the wrong number changed. I assumed he changed the type to type 5, and I think that was the intention?

2021-05-09, 10:20	#1
ric Jul 2004 Milan, Ita 2·7·13 Posts	Malformed PRP-CF assignment from server Got assigned Code: PRP=<aid>,1,2,3215747,-1,99,0,3,1,"4457025343,185822885311153245017,185822885311153245017" from Primenet, and mprime complained with Code: 185822885311153245017 does not divide exponent which is clearly not true, and discarded the worktodo line. Double-checked the assignment line from my assignments list, and realized that in there the recent factor was contained twice (hence the warning); wandering whether this depends from it being reported twice to the server. Removed the duplicate from the assignment line, changed the '99,0,3,1' to '99,0,5,1', now happily crunching. I'm pretty confident this is a very corner-case, so I'm leaving this note mainly as a self-reminder.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Gerbicz Error Checking - Originally: Would a "rack server" run the same as a desktop server	EdH	Hardware	16	2021-03-29 16:43
Use of a virtual private server (VPS) - normally a web server	drkirkby	Hardware	11	2021-03-15 09:11
Server handling of Cert assignment for the next MP	Uncwilly	PrimeNet	0	2020-09-02 21:35
v4 server's assignment report	Prime95	Data	7	2008-11-21 22:47
New Server Hardware and price quotes, Funding the server	Angular	PrimeNet	32	2002-12-09 01:12

2021-05-09, 11:00	#2
ATH Einyen Dec 2003 Denmark 3⁵×13 Posts	Yes, this is a known error with unknown cause: https://mersenneforum.org/showpost.p...&postcount=258 https://mersenneforum.org/showpost.p...&postcount=259

2021-05-20, 21:59	#4
ATH Einyen Dec 2003 Denmark 3⁵×13 Posts	PRP-CF are usually run with type 5, because then if a new factor is found, only 1 more test is needed, because the residue will be the same with the new factor. https://mersenneforum.org/showthread.php?t=25841

2021-05-21, 08:34	#6
kruoli "Oliver" Sep 2017 Porta Westfalica, DE 7²×11 Posts	Yes, the test does show up as using base 5 and type 1 on the results page: 3215747.