Expired assignment

kracker · 2016-01-03, 02:10

I'm not sure if I did something wrong.. but yesterday, I reserved a few P-1 assignments from manual testing, and now after I submitted one it's shown as "expired". Is that normal?
http://www.mersenne.org/report_expon...5171829&full=1

Madpoo · 2016-01-03, 23:35

Quote:

Originally Posted by kracker

I'm not sure if I did something wrong.. but yesterday, I reserved a few P-1 assignments from manual testing, and now after I submitted one it's shown as "expired". Is that normal?
http://www.mersenne.org/report_expon...5171829&full=1

If it was checked in without the assignment ID, it might have expired the assignment instead of marking it complete. I'm not sure if P-1 works that way or not... it might... so that's just a guess.

LaurV · 2016-01-04, 02:58

We kinda had the same problems in the past, some assignments of ourselves became expired, instead of "accomplished", when we reported the results. We complained in the forum but no action, the bothering part is that one of them appears in the CPU list, which is even stranger. Let me look for it.

Ok, is this one. When I open the CPUs list, it shows me that I still have this assignment, but in the assignment list, I don't have it. The work was completed a year ago, although with higher limits than the PrimeNet expected, maybe this is the cause? However the work was reported with the right key.

Syntony · 2016-01-06, 01:01

Something similar happened to me a couple of weeks back, when I released a P-1 assignment through Prime95 itself - the Prime95 Communication thread reported an error, saying that the P-1 test wasn't assigned to me. The weird part was that the assignment key reported in the error message bore no relation to the key of the assignment I was releasing. When I looked at the exponent I'd just released, it was indeed still shown as assigned to me, though it had disappeared from my GIMPS Assignment Details report...

At the time I just assumed that the key had been corrupted on the comms link, but maybe not... In any case I recreated the assignment in my worktodo.txt file using prime.log to recover the assignment key and freed it again, which resulted in the assignment being cleared in the normal way.

Perhaps there really is a bug in GIMPS which (occasionally) causes the key to be corrupted during processing?

Madpoo · 2016-01-06, 03:44

Quote:

Originally Posted by Syntony

Perhaps there really is a bug in GIMPS which (occasionally) causes the key to be corrupted during processing?

Hmm... maybe. Might be specific to P-1 even.

petrw1 · 2016-01-19, 05:03

Code:

Exponent Work Type Stage,% Age (days) ETA (days) Assigned Last Update Next Update  Estimated Completion  Userid 
66587777 LL LL, 99.90% 86 -60 2015-10-25 2015-11-20 2015-11-21 2015-11-20 ANONYMOUS 
66589511 LL LL, 99.80% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66591037 LL LL, 99.60% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66587839 LL LL, 99.20% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66593491 LL LL, 98.70% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS

rudi_m · 2016-01-20, 10:06

I have also a similar problem with double assignment.

About one week ago one of my machines assigned M332301763 (worktype 102 - World record primality tests).

It completed P-1 stage 1 and 2 and I've got NF-PM1 credits for that. While P-1 it was logging these warnings:
" WARNING: Estimated completion date is more than one year away. The assignment may be reassigned to another user after one year".

After finishing P-1 I moved this LL job to another faster machine. But it looks odd.

My assignment page shows workype PM1-R and not LL.
Also assignment page "worktodo entries" shows "UNKNOWN WORKTYPE "95".
AND somebody else also assigned this exponent:

Code:

   Assigned     2016-01-11    rudimeier    PM1-R	
   Assigned     2016-01-19    yekose       LL

http://www.mersenne.org/report_expon...2301763&full=1

retina · 2016-01-20, 14:33

Quote:

Originally Posted by petrw1

Code:

Exponent Work Type Stage,% Age (days) ETA (days) Assigned Last Update Next Update  Estimated Completion  Userid 
66587777 LL LL, 99.90% 86 -60 2015-10-25 2015-11-20 2015-11-21 2015-11-20 ANONYMOUS 
66589511 LL LL, 99.80% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66591037 LL LL, 99.60% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66587839 LL LL, 99.20% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS 
66593491 LL LL, 98.70% 86 -59 2015-10-25 2015-11-20 2015-11-21 2015-11-21 ANONYMOUS

This problem might go away if the P95 client would report the result immediately upon completion of the test. No doubt the user(s) saw the exponent finish with a non-zero result(s) and then closed P95 (and perhaps also deleted it in disappointment).

Prime95 · 2016-01-20, 15:20

Quote:

Originally Posted by rudi_m

About one week ago one of my machines assigned M332301763 (worktype 102 - World record primality tests).

My assignment page shows workype PM1-R and not LL.
Also assignment page "worktodo entries" shows "UNKNOWN WORKTYPE "95".
AND somebody else also assigned this exponent

That is a head scratcher. First of all, worktype 102 should get you an exponent just above 74.2M, not 332M.

I'll have to dig around to see if I can figure out a way your assignment got turned into an invalid worktype.

Going forward, do you really want to test a 100 million digit number? They take a long, long time. If you do continue, the other assignee will probably not finish it. If you try to get a new worktype 102 exponent, don't start it if you get another 100 million digit number.

rudi_m · 2016-01-20, 20:07

Quote:

Originally Posted by Prime95

That is a head scratcher. First of all, worktype 102 should get you an exponent just above 74.2M, not 332M.

Ah sorry, here my memory was wrong! Surely I had set worktype 104!

BTW I've got the exponent while I was in the menu "./mprime -m". After I've got it I switched back WorkPreference=101 and started "./mprime -d" to finish PM1.

Quote:

Originally Posted by Prime95

Going forward, do you really want to test a 100 million digit number? They take a long, long time. If you do continue, the other assignee will probably not finish it. If you try to get a new worktype 102 exponent, don't start it if you get another 100 million digit number.

Yes, I want to finish. It will take only 120 days on this i5-6500 (3 threads). If the other one will finish too then I would be happy about the double check. But only one of us could win the medal ;(

Brian-E · 2016-01-22, 10:09

Quote:

Originally Posted by rudi_m

[...]Yes, I want to finish. It will take only 120 days on this i5-6500 (3 threads).[...]

I take longer than that to do my current double-check of a 42M exponent! You people with your modern hardware make my contribution laughable.

2016-01-03, 02:10	#1
kracker "Mr. Meeseeks" Jan 2012 California, USA 100010000111₂ Posts	Expired assignment I'm not sure if I did something wrong.. but yesterday, I reserved a few P-1 assignments from manual testing, and now after I submitted one it's shown as "expired". Is that normal? http://www.mersenne.org/report_expon...5171829&full=1

2016-01-20, 10:06	#7
rudi_m Jul 2005 266₈ Posts	I have also a similar problem with double assignment. About one week ago one of my machines assigned M332301763 (worktype 102 - World record primality tests). It completed P-1 stage 1 and 2 and I've got NF-PM1 credits for that. While P-1 it was logging these warnings: " WARNING: Estimated completion date is more than one year away. The assignment may be reassigned to another user after one year". After finishing P-1 I moved this LL job to another faster machine. But it looks odd. My assignment page shows workype PM1-R and not LL. Also assignment page "worktodo entries" shows "UNKNOWN WORKTYPE "95". AND somebody else also assigned this exponent: Code: Assigned 2016-01-11 rudimeier PM1-R Assigned 2016-01-19 yekose LL http://www.mersenne.org/report_expon...2301763&full=1

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
was P-1 expired?	dh1	PrimeNet	9	2016-09-04 14:18
Is it OK to pick up expired exponents?	patrik	PrimeNet	9	2014-04-09 23:30
Exponent expired?	Chuck	PrimeNet	4	2014-02-11 03:49
Question regarding expired LL exponents	E_tron	PrimeNet	2	2010-11-09 18:08
Expired exponents	justinsane	PrimeNet	8	2004-03-24 06:22

2016-01-04, 02:58	#3
LaurV Romulan Interpreter "name field" Jun 2011 Thailand 2⁴·643 Posts	We kinda had the same problems in the past, some assignments of ourselves became expired, instead of "accomplished", when we reported the results. We complained in the forum but no action, the bothering part is that one of them appears in the CPU list, which is even stranger. Let me look for it. Ok, is this one. When I open the CPUs list, it shows me that I still have this assignment, but in the assignment list, I don't have it. The work was completed a year ago, although with higher limits than the PrimeNet expected, maybe this is the cause? However the work was reported with the right key.

2016-01-06, 01:01	#4
Syntony "Tony" Sep 2014 London, UK 2²·23 Posts	Something similar happened to me a couple of weeks back, when I released a P-1 assignment through Prime95 itself - the Prime95 Communication thread reported an error, saying that the P-1 test wasn't assigned to me. The weird part was that the assignment key reported in the error message bore no relation to the key of the assignment I was releasing. When I looked at the exponent I'd just released, it was indeed still shown as assigned to me, though it had disappeared from my GIMPS Assignment Details report... At the time I just assumed that the key had been corrupted on the comms link, but maybe not... In any case I recreated the assignment in my worktodo.txt file using prime.log to recover the assignment key and freed it again, which resulted in the assignment being cleared in the normal way. Perhaps there really is a bug in GIMPS which (occasionally) causes the key to be corrupted during processing?