mersenneforum.org  

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

Reply
 
Thread Tools
Old 2021-05-27, 16:03   #2168
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

3·7·167 Posts
Default

Quote:
Originally Posted by axn View Post
Second factor is not properly recorded: https://www.mersenne.org/report_expo...3731081&full=1
Can you confirm if this was submitted to the server automatically (via API) or through the manual results form? I tried resubmitting your result on the manual form and it appeared to record both factors correctly, which would point to the error being in the API side of things, if you could confirm that's how it was submitted?
James Heinrich is offline   Reply With Quote
Old 2021-05-27, 17:34   #2169
axn
 
axn's Avatar
 
Jun 2003

515110 Posts
Default

Manual.
axn is offline   Reply With Quote
Old 2021-05-27, 18:05   #2170
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

581310 Posts
Default

Quote:
Originally Posted by Prime95 View Post
Question: Do the assignments actually expire in 7 days or are they just reported as cat 0's expiring in 7 days?
To follow up on that, and part of post 2164, although the manual assignment was a PRP requested as a DC, https://www.mersenne.org/report_expo...4607561&full=1 went from 7 days down to 1 day to expiration in the assignments report, to 0, to 22 days, before I reported in the result, being shown as a cat 0 PRP the whole time. I have no way of checking via the web pages whether there was any admin intervention contributing to that increase to 22, or if the server somehow did that on its own.

Last fiddled with by kriesel on 2021-05-27 at 18:07
kriesel is offline   Reply With Quote
Old 2021-05-27, 18:09   #2171
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

3·7·167 Posts
Default

Quote:
Originally Posted by axn View Post
Manual.
I have no immediate explanation as to why your submission (apparently) only processed the first factor but no the second whereas mine processed both. It might have been helpful to see the entire transaction (e.g. screenshot) rather than copy-paste of selected output lines. My best guess at this point is that something about the "you already submitted this result" error message you quoted is triggering a SQL transaction rollback and "forgetting" the second factor, but that doesn't explain how or why my submission worked and yours didn't.

If/when someone else finds a new composite factor (P-1 or otherwise) please double-check that all prime factors were correctly recorded, and if they did not I'd appreciate a screenshot (or save-as-html) of the manual result form screen output to see if I can narrow down the source of the problem.

Last fiddled with by James Heinrich on 2021-05-27 at 18:10
James Heinrich is offline   Reply With Quote
Old 2021-05-27, 20:25   #2172
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

3,313 Posts
Default

Quote:
Originally Posted by kriesel View Post
...went from 7 days down to 1 day to expiration in the assignments report, to 0, to 22 days, before I reported in the result, being shown as a cat 0 PRP the whole time. I have no way of checking via the web pages whether there was any admin intervention contributing to that increase to 22, or if the server somehow did that on its own.
Assignments start out with a 7 day countdown, until at least some progress or check-in has occurred. Even if the progress being checked in is 0.0%, it'll count as "I started this" and then the regular thresholds apply.

So maybe that's what happened? It was assigned, but zero progress, and the 7 day timer was going. Then on that 7th day, you reported some progress so now the countdown assumed you had the full 30 or whatever the case. Could be.
Madpoo is offline   Reply With Quote
Old 2021-05-27, 20:57   #2173
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

5,813 Posts
Default

Quote:
Originally Posted by Madpoo View Post
Assignments start out with a 7 day countdown, until at least some progress or check-in has occurred. Even if the progress being checked in is 0.0%, it'll count as "I started this" and then the regular thresholds apply.

So maybe that's what happened? It was assigned, but zero progress, and the 7 day timer was going. Then on that 7th day, you reported some progress so now the countdown assumed you had the full 30 or whatever the case. Could be.
Thanks for responding. Sure, that happens with PrimeNet-connected prime95 & mprime, which is not the case I was writing about. I don't understand how checkin could relate to a totally manual assignment on gpuowl on a GPU. How could it? There is no communication link between the GPU application and the server, no PrimeNet API involved. There is no way even with a user's manual intervention, to check in progress state of a PRP or LL test for a manual GPU run at the web server. I've requested that capability repeatedly, for years, and IIRC no one credible has ever responded, here's how to do that, or we'll try to create that. I'm not even using gpuowl primenet.py, which as I understand it does assignment and result transmissions only, not any progress updating or checkin during the course of running a primality test. In this case, I intentionally waited for a day for the 0 days left to go away, to see what came next, expiration, blank days to expiration, what. So it is not a case of having duplicated the assignment onto a prime95 session briefly to stall expiration, but the opposite case, of waiting and watching for expiration.

If you know of a way to do progress reporting for GPU apps during a primality test, please enlighten us how.

FYI I have several manual 8xM PRP DC assignments on two Radeon VII GPUs, indicated as expiring in 3 days. Most should clear fine before hitting 0 days or expiring, but the last one will be close.

Last fiddled with by kriesel on 2021-05-27 at 21:04
kriesel is offline   Reply With Quote
Old 2021-05-27, 22:01   #2174
chalsall
If I May
 
chalsall's Avatar
 
"Chris Halsall"
Sep 2002
Barbados

132×59 Posts
Default

Quote:
Originally Posted by kriesel View Post
If you know of a way to do progress reporting for GPU apps during a primality test, please enlighten us how.
Not meaning to speak for Madpoo, but the solution space is trivial.

While your serious kit does the work, have proxies reserve and update on progress.

This is basically what I do with GPU72. Seems to scale fairly well.
chalsall is offline   Reply With Quote
Old 2021-05-28, 02:04   #2175
axn
 
axn's Avatar
 
Jun 2003

515110 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
I have no immediate explanation as to why your submission (apparently) only processed the first factor but no the second whereas mine processed both. It might have been helpful to see the entire transaction (e.g. screenshot) rather than copy-paste of selected output lines.
The quoted text is verbatim the output of processing of the affected exponent. There were couple of other submissions before & after. If you can't narrow down the issue from the text, I don't think screenshot would have helped.
axn is offline   Reply With Quote
Old 2021-05-28, 03:37   #2176
slandrum
 
Jan 2021
California

3448 Posts
Default

Quote:
Originally Posted by kriesel View Post
To follow up on that, and part of post 2164, although the manual assignment was a PRP requested as a DC, https://www.mersenne.org/report_expo...4607561&full=1 went from 7 days down to 1 day to expiration in the assignments report, to 0, to 22 days, before I reported in the result, being shown as a cat 0 PRP the whole time. I have no way of checking via the web pages whether there was any admin intervention contributing to that increase to 22, or if the server somehow did that on its own.
I believe what happened here is that manual assignments are not held to the "you must start the work in x days" rule, probably because there's no (easy) way to report progress from some of the applications. In theory you aren't supposed to get cat 0 or cat 1 manual assignments, so the server probably reports the "work hasn't started" expiration on them even though it doesn't apply. The real expiration was the "assignment is too old" which you started seeing after the "you haven't started the assignment" expiration passed.

Last fiddled with by slandrum on 2021-05-28 at 03:43
slandrum is offline   Reply With Quote
Old 2021-05-28, 04:10   #2177
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

7,639 Posts
Default

Quote:
Originally Posted by slandrum View Post
I believe what happened here is that manual assignments are not held to the "you must start the work in x days" rule, probably because there's no (easy) way to report progress from some of the applications.
That is true. Ken's bug report then changes to "why does the assignment's page apply the will expire in x days if not started rule to manual assignments?"
Prime95 is online now   Reply With Quote
Old 2021-05-28, 08:33   #2178
S485122
 
S485122's Avatar
 
"Jacob"
Sep 2006
Brussels, Belgium

2·32·97 Posts
Default

Quote:
Originally Posted by kriesel View Post
...
There is no way even with a user's manual intervention, to check in progress state of a PRP or LL test for a manual GPU run at the web server. I've requested that capability repeatedly, for years, and IIRC no one credible has ever responded, here's how to do that, or we'll try to create that.
...
If you know of a way to do progress reporting for GPU apps during a primality test, please enlighten us how.
One way to solve this problem is to limit the time by which manual assignments can be extended. It could even be refined by modulating the length of the extension by category of exponent.
An manual extension of the assignment could be counted as an update..

The lower category assignments (manual or not) must be shepherded by their owners ;-)

Jacob
S485122 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 01:44.


Thu Oct 28 01:44:42 UTC 2021 up 96 days, 20:13, 0 users, load averages: 2.88, 2.94, 2.96

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.