mersenneforum.org  

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

Reply
 
Thread Tools
Old 2021-05-28, 16:28   #2179
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

132658 Posts
Default

I think the issue is that a PRP DC manual assignment of 84-90M exponent, which it seems to me should be treated as a Cat FOUR DC, since today's upper limit for cat3 DC is ~63M, appears to be subjected to cat ZERO first test update or expire requirements.

This comes from PRP and PRPDC assignments being both treated as if they were first-test assignments.

Cat 0 first test LL upper limit now is ~103.75M. So incorrectly applying first-test rules to manual checkouts for DC, put them in Cat 0. Even though they were obtained as DC, and manual assignment is not supposed to issue cat 0 or 1. ("The server oughta know better!" :)

And I note that the assignment rules page doesn't say anything applies or differs for PRP. The page contains zero occurrences of "PRP".

One of the points of getting manual reservations ~30M above the DC milestone level, is to comfortably avoid expiration issues for slow running gpus or deep queues.
That's defeated when the server handles DC cat 4 as first-test cat 0.

Manual extension is not effective in changing expiration indications on a user's assignments page. It shows extension on the manual extension page afterward, but it gets reset by the server later somehow.
kriesel is offline   Reply With Quote
Old 2021-05-29, 07:34   #2180
Aramis Wyler
 
Aramis Wyler's Avatar
 
"Bill Staffen"
Jan 2013
Pittsburgh, PA, USA

42410 Posts
Default

Was something done on this? My time to expire numbers all shot dramatically upwards, which is comforting since they had all been red. That's across 2 different cpu architectures and a gpu.


I appreciate the improvement. I've never had an exponent actually expire from this, but a lot of DCs on my slow work pc get done with -12 or even -19 days remaining sometimes.
Aramis Wyler is offline   Reply With Quote
Old 2021-05-29, 20:12   #2181
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

5,813 Posts
Default no P-1

There's a report with example, of the server giving out PRP assignments, to mprime via PrimeNet API, for exponents with no prior P-1 performed, in a form that prevents P-1 factoring before PRP, where available allowed RAM is very abundant. See https://mersenneforum.org/showthread.php?t=26849 for more info.
kriesel is offline   Reply With Quote
Old 2021-05-31, 01:35   #2182
axn
 
axn's Avatar
 
Jun 2003

10100000111112 Posts
Default Fresh off the oven

Composite factor submission error (manual form). Screen shot attached.

Code:
Splitting composite factor 10913007652382289864756170952954868628781851495647 into:
* 1101799602906346391183783
* 9904711912761418810995209
processing: P-1 factor 1101799602906346391183783 for M3904631 (B1=14,000,000, B2=700,000,000) (79.866 bits)
CPU credit is 8.1907 GHz-days.
processing: P-1 factor 9904711912761418810995209 for M3904631 (B1=14,000,000, B2=700,000,000) (83.034 bits)
Error code: 40, error text: You have already sent in this P-1 result for M3904631
Attached Thumbnails
Click image for larger version

Name:	composite_factor_error.png
Views:	40
Size:	148.8 KB
ID:	24988  
axn is offline   Reply With Quote
Old 2021-05-31, 01:50   #2183
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

3·7·167 Posts
Default

Thank you, this is helpful.

edit: The problem is (probably) fixed now, we'll need to wait for discovery of another composite factor to confirm fix.

Last fiddled with by James Heinrich on 2021-05-31 at 03:45
James Heinrich is online now   Reply With Quote
Old 2021-06-01, 22:29   #2184
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

10110101101012 Posts
Default The server lies to us on the personal assignments page

It misrepresents manual assignments' expiration dates. And the day after reaching zero in the misrepresentation, changes to a new misrepresentation. Manual extension by the user has no apparent effect on that behavior. Manual extension changes the "last updated" date but has no effect on the displayed expiration date. And it represents PRPDC assignments it issued, as PRP first tests. Following is tracking over time of how a specific set of exponents reserved manually as PRP DC were displayed from day to day. Exponents drop out of the report when a result is manually submitted. Some results were intentionally delayed slightly to see what the server would do next.
Code:
Focus on manual PRP DC assignments from web page, gpuowl, no updating of progress possible

An earlier snapshot; some of which were reported in with results lines manually, then proof files dropped into a primenet-connected  running prime95 instance.
"Cpu name"   "core"     exponent    work-type cat   original  (no stage  assigned     last updated  (no est.  (no days  expires
                          now     cat    or %)                    complete)  to go)   (days)
Manual testing    1    84607561      PRP    0    0        2021-05-18    2021-05-18            1
Manual testing    1    84944921      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    84969113      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    85141577      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    85427897      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    86042767      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    87051017      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    87215537      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    88102523      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    88391669      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    89023807      PRP    0    0        2021-05-23    2021-05-23            6
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            6


May 27 1616 US CDT (UTC -0500), 2116 UTC
"Cpu name"   "core"     exponent    work-type cat   original  (no stage  assigned     last updated  (no est.  (no days  expires
                          now     cat    or %)                    complete)  to go)   (days)
Manual testing    1    85427897      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    86042767      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    87215537      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    88102523      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    88391669      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    89023807      PRP    0    0        2021-05-23    2021-05-23            3
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            3
(Exponent data copy/pasted to notepad from https://www.mersenne.org/workload/ for kriesel; header typed in)


May 28 ~1132 US CDT, ~1632 UTC
"Cpu name"   "core"     exponent    work-type cat   original  (no stage  assigned     last updated  (no est.  (no days  expires
                          now     cat    or %)                    complete)  to go)   (days)
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    88102523      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    88391669      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    89023807      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            2


Okay, lets try manual extension of the first 86961583 and second 88102523, then recheck. On the manual extensions page, afterward:
"Cpu name"   "core"      number     work type  stage   age    days   expires  estimated   next        updated
                             %    days    to go    days  completion     update     
Manual testing    0    M86961583     PRP    0.00%    5        235        2022-01-08    2021-05-28
Manual testing    0    M88102523     PRP    0.00%    5        235        2022-01-08    2021-05-28
Manual testing    0    M88391669     PRP    0.00%    5        175        2021-11-09    2021-05-23
Manual testing    0    M89023807     PRP    0.00%    5        175        2021-11-09    2021-05-23
Manual testing    0    M89877037     PRP    0.00%    5        175        2021-11-09    2021-05-23


Now a fresh load of the user assignments page shows:
"Cpu name"   "core"     exponent    work-type cat   original  (no stage  assigned     last updated  (no est.  (no days  expires
                          now     cat    or %)                    complete)  to go)   (days)
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-28            2
Manual testing    1    88102523      PRP    0    0        2021-05-23    2021-05-28            2
Manual testing    1    88391669      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    89023807      PRP    0    0        2021-05-23    2021-05-23            2
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            2

I wonder what tomorrow will show. My guess is any survivors not reported as completed results will show expire in 1 day.

May 29 1245 US CDT, 1745 UTC
"Cpu name"   "core"     exponent    work-type cat   original  (no stage  assigned     last updated  (no est.  (no days  expires
                          now     cat    or %)                    complete)  to go)   (days)
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-28            1
Manual testing    1    89023807      PRP    0    0        2021-05-23    2021-05-23            1
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            1

May 30 0805 US CDT, 1305 UTC
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-28            0
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            0

May 31 0815 US CDT, 1315 UTC
Manual testing    1    86961583      PRP    0    0        2021-05-23    2021-05-28            22
Manual testing    1    89877037      PRP    0    0        2021-05-23    2021-05-23            22

(end)

Last fiddled with by kriesel on 2021-06-01 at 22:31
kriesel is offline   Reply With Quote
Old 2021-06-05, 06:27   #2185
Happy5214
 
Happy5214's Avatar
 
"Alexander"
Nov 2008
The Alamo City

3×11×23 Posts
Default

I resubmitted my entire results.json.txt from my manual results mprime instance, just to make sure I didn't forget anything. It accepted one line as not being a duplicate, for a stage 1-only P-1 on 2273041, but it I think that should have been marked as a duplicate too (there are now three P-1 lines, two stage 1-only with the same B1 and one with both stages). Can you delete the duplicate and figure out why that JSON line wasn't in the system already?
Happy5214 is offline   Reply With Quote
Old 2021-06-05, 17:46   #2186
Dobri
 
May 2018

3078 Posts
Default M103313627

The same server problem that occurred in March of this year in the 101-xxxxxx range is now observed for the current lowest untested exponent 103313627 in the 103-xxxxxx range.
It remains unassigned since mid-May, see the attached images.

P.S. Note that said server problem in the 101-xxxxxx range still needs a closure in the Work Distribution Map, because 54316 < 35141 + 14718 + 4458.
54316 | 35141 14718 4458
Attached Thumbnails
Click image for larger version

Name:	M103313627.jpg
Views:	44
Size:	374.9 KB
ID:	25015   Click image for larger version

Name:	M103313627b.jpg
Views:	43
Size:	487.7 KB
ID:	25016  
Dobri is offline   Reply With Quote
Old 2021-06-08, 07:43   #2187
drkirkby
 
"David Kirkby"
Jan 2021
Althorne, Essex, UK

6778 Posts
Default

Perhaps I am wrong, but is looks like the same person (Ben Delo) who did the PRP test, has also verified the proof on these exponents.

https://www.mersenne.org/report_expo...2109703&full=1
https://www.mersenne.org/report_expo...2110257&full=1
https://www.mersenne.org/report_expo...2110971&full=1
https://www.mersenne.org/report_expo...2111397&full=1
https://www.mersenne.org/report_expo...2111431&full=1
https://www.mersenne.org/report_expo...2111479&full=1

Maybe the server can't keep up with the work that Ben does!

I guess if the server is making an error and occasionally sending the PRP test and the certificate to the same person, this would more likely be spotted with Ben, due to the sheer amount of work he does. Does the server just send the certificate work to a random person, without checking if they did the PRP test too? In general that is unlikely to result in the same person getting both, but since Ben is doing 10x more first-time PRP tests than anyone else, it is more likely to show up on his account than anyone elses.

Last fiddled with by drkirkby on 2021-06-08 at 08:24
drkirkby is offline   Reply With Quote
Old 2021-06-08, 09:01   #2188
kruoli
 
kruoli's Avatar
 
"Oliver"
Sep 2017
Porta Westfalica, DE

23·5·17 Posts
Default

It is okay for the same person doing the PRP test and the proof verification, since the PrimeNet server will do some work in between which makes sure the following proof will be practically tamper proof.
kruoli is offline   Reply With Quote
Old 2021-06-08, 10:37   #2189
drkirkby
 
"David Kirkby"
Jan 2021
Althorne, Essex, UK

3·149 Posts
Default

Quote:
Originally Posted by kruoli View Post
It is okay for the same person doing the PRP test and the proof verification, since the PrimeNet server will do some work in between which makes sure the following proof will be practically tamper proof.
Okay, that makes sense. Thank you.
drkirkby is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Official "Faits erronés dans de belles-lettres" thread ewmayer Lounge 39 2015-05-19 01:08
Official "all-Greek-to-me Fiction Literature and Cinema" Thread ewmayer Science & Technology 41 2014-04-16 11:54
Official "Lasciate ogne speranza" whinge-thread cheesehead Soap Box 56 2013-06-29 01:42
Official "Ernst is a deceiving bully and George is a meanie" thread cheesehead Soap Box 61 2013-06-11 04:30
Official "String copy Statement Considered Harmful" thread Dubslow Programming 19 2012-05-31 17:49

All times are UTC. The time now is 01:10.


Thu Oct 28 01:10:18 UTC 2021 up 96 days, 19:39, 0 users, load averages: 2.49, 2.75, 2.79

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.