climateprediction.net home page
Server State Over, but wu is in progress!

Server State Over, but wu is in progress!

Message boards : Number crunching : Server State Over, but wu is in progress!
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Profile old_user633

Send message
Joined: 9 Aug 04
Posts: 33
Credit: 168,775
RAC: 0
Message 19094 - Posted: 9 Jan 2006, 19:03:49 UTC

I have 2 wu\'s currently in progress but the result listing page (link below) shows that the results are Over. Since they are both in progress and not timed out and they are not completed is this a bug? What implicatons does it have (if any)?

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?userid=633

Over The result has been sent to a host and either it has timed out or the host has reported its completion.

Live long and crunch.

Paul
(S@H1 8888)
<img border="0" height="80" src="http://31411.sah.sig.boinc.dk?188"></a>BOINC/SAH BETA<img src="http://mysite.freeserve.com/gifcollection/img52.gif">
ID: 19094 · Report as offensive     Reply Quote
Profile Honza
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 390
Credit: 2,475,242
RAC: 0
Message 19097 - Posted: 9 Jan 2006, 19:30:38 UTC

See Carl\'s reply here - http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=3746#18635
btw, the link doesn\'t work; should have been provided to your host like
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=214525
<i>phpBB forum for CPDN, all are </i><a href="http://www.climateprediction.net/board">invited</a>
ID: 19097 · Report as offensive     Reply Quote
Profile old_user160

Send message
Joined: 5 Aug 04
Posts: 10
Credit: 106,631
RAC: 0
Message 19099 - Posted: 9 Jan 2006, 21:21:01 UTC - in response to Message 19094.  
Last modified: 9 Jan 2006, 21:23:35 UTC

I\'ve been attempting to contact the people on the BOINC BETA (Seti Beta) boards about implementing a method for the client to request more time for work if it is nearing it\'s expiration date for reasons similar to this. (although maybe not for this particular reason!)

HOWEVER - Everyone keeps calling me crazy, claiming that \"computers would never have enough work cached locally to pass the deadlines.. \" personally I wonder about what has happened to the calibur of beta testers in the past couple years...
ID: 19099 · Report as offensive     Reply Quote
Profile old_user633

Send message
Joined: 9 Aug 04
Posts: 33
Credit: 168,775
RAC: 0
Message 19115 - Posted: 10 Jan 2006, 5:15:18 UTC

Doh! Thanks Honaza, I should\'ve known.

So are there any implications for continued crunching of these wu\'s? I\'d hate to be crunching for another 1000hrs for nothing and I just don\'t mean the credits. I\'d rather be crunching for science needs and not just the credit (which pains me to say since I am a credit whore).
ID: 19115 · Report as offensive     Reply Quote
Profile old_user5994

Send message
Joined: 31 Aug 04
Posts: 239
Credit: 2,933,299
RAC: 0
Message 19129 - Posted: 10 Jan 2006, 9:37:44 UTC - in response to Message 19099.  

HOWEVER - Everyone keeps calling me crazy, claiming that \"computers would never have enough work cached locally to pass the deadlines.. \" personally I wonder about what has happened to the calibur of beta testers in the past couple years...

Probably lost due to the non-responsiveness of development teams ...

Heck, I pointed to a bug in the feeder code that may be the cause of the feeder problems experienced at Predictor@Home and SETI@Home (Rosetta@Home?) with the feeder not having enough work on hand... that was 5 days ago ... not even an acknowledgement (yet) ...

Since I have never been able to \"think\" in C (since Java is C in drag, same there); not sure if my suggested change is viable.
ID: 19129 · Report as offensive     Reply Quote
Profile Honza
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 390
Credit: 2,475,242
RAC: 0
Message 19139 - Posted: 10 Jan 2006, 13:23:58 UTC - in response to Message 19115.  

So are there any implications for continued crunching of these wu\'s? [...] I\'d rather be crunching for science needs and not just the credit (which pains me to say since I am a credit whore).
I think it is still valuable to run those runs.
http://www.climateprediction.net/board/viewtopic.php?p=32429#32429

But running whole model is more (most) valuable and even makes more sense to those looking at results of all phases.
<i>phpBB forum for CPDN, all are </i><a href="http://www.climateprediction.net/board">invited</a>
ID: 19139 · Report as offensive     Reply Quote
Profile Keck_Komputers
Avatar

Send message
Joined: 5 Aug 04
Posts: 426
Credit: 2,426,069
RAC: 0
Message 19158 - Posted: 10 Jan 2006, 20:53:29 UTC - in response to Message 19099.  

I\'ve been attempting to contact the people on the BOINC BETA (Seti Beta) boards about implementing a method for the client to request more time for work if it is nearing it\'s expiration date for reasons similar to this. (although maybe not for this particular reason!)

HOWEVER - Everyone keeps calling me crazy, claiming that \"computers would never have enough work cached locally to pass the deadlines.. \" personally I wonder about what has happened to the calibur of beta testers in the past couple years...

This idea has been batted around for years. It\'s not that you are crazy, but there hasn\'t been a means to do it. I am hopefull that if we can get enough of the server side calculations done on the client that this can be implemented. However don\'t look for it anytime soon.
BOINC WIKI

BOINCing since 2002/12/8
ID: 19158 · Report as offensive     Reply Quote
Profile old_user633

Send message
Joined: 9 Aug 04
Posts: 33
Credit: 168,775
RAC: 0
Message 19163 - Posted: 11 Jan 2006, 2:13:39 UTC
Last modified: 11 Jan 2006, 2:14:57 UTC

Another observation is that the cpu time on the results page (link below) is incorrect when compared to the trickle information page (see top result and the third result).

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=214525

The 3rd result is also not going to make it by its deadline in 13 days (due to my earlier intervention and then a failed power supply which took 2 weeks to get a replacement under warranty). I want to continue it to it\'s end, is it worth it? It is currently at 65% complete and by its deadline it will be at approx 85% complete. By the time it completes it will be about 10 days passed the deadline. Will it still be useful for the project (Honza appears to say yes)? Will a new result be generated and sent out since this one is not completed in time?

Live long and crunch.

Paul.
ID: 19163 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 19164 - Posted: 11 Jan 2006, 2:50:10 UTC

Deadlines aren\'t enforced. Provided that the model trickles regularly to let the server know that it\'s still alive, the data will be accepted.
Of course, \"regularly\" is open to interpretation, so let\'s say, every 2-3 days max. More often is better.

ID: 19164 · Report as offensive     Reply Quote
Profile old_user633

Send message
Joined: 9 Aug 04
Posts: 33
Credit: 168,775
RAC: 0
Message 19198 - Posted: 12 Jan 2006, 8:35:12 UTC

Thanks Les. This wu will be crunching 24/7 on 1 of 2 HT cpu\'s. It trickles about every 12.5hrs, so it sounds like it will be OK.

Live long and crunch.

Paul.
ID: 19198 · Report as offensive     Reply Quote
Profile Ananas
Volunteer moderator

Send message
Joined: 31 Oct 04
Posts: 336
Credit: 3,316,482
RAC: 0
Message 19200 - Posted: 12 Jan 2006, 9:28:34 UTC
Last modified: 12 Jan 2006, 9:30:42 UTC

Not related the question but maybe good to know (I have tested this only with BOINC 4.19 though) :

On a machine with 2 tasks (HT or dual CPU), you can force BOINC to run 1 CPDN task all the time and use the other CPU for different projects by giving CPDN less than 50% ressource share but a really high (e.g. a few 100k) &lt;debt&gt; value in client_state.xml

Don\'t forget to shutdown BOINC while you modify client_state.xml, it\'s really important to do that or you might loose work. And be careful to save it as a plain text, if your favourite editor is something like wordpad. Better use a plain text editor to reduce the risk of messing things up.
ID: 19200 · Report as offensive     Reply Quote
Profile old_user5994

Send message
Joined: 31 Aug 04
Posts: 239
Credit: 2,933,299
RAC: 0
Message 19201 - Posted: 12 Jan 2006, 11:14:44 UTC

Be sure to set the project to no new work ... or you are likely to get another work unit to \"help\" ... :)

In my case, I download a work unit, then set no new work until the work unit has been reported, rinse and repeat. This way I do not have 23 CPDN work units hanging about ... of course I have traditionally give CPDN a large slice of my computer\'s time also ...
ID: 19201 · Report as offensive     Reply Quote
Profile Ananas
Volunteer moderator

Send message
Joined: 31 Oct 04
Posts: 336
Credit: 3,316,482
RAC: 0
Message 19208 - Posted: 12 Jan 2006, 16:18:42 UTC

4.19 i.e. no \"no new work\" option, this feature came later

But the 4.19 scheduler is less \"smart\" so it\'s easier to trick, at 50% resource share it did download a second model but with less than 50% it seems to be happy with one model.
ID: 19208 · Report as offensive     Reply Quote
Profile old_user5994

Send message
Joined: 31 Aug 04
Posts: 239
Credit: 2,933,299
RAC: 0
Message 19209 - Posted: 12 Jan 2006, 17:19:35 UTC

Lord only knows why you would want to be that far out of date... but you can acheive the same effect by playing with your disk space numbers ... just a tip if you persist in the madness ... :)
ID: 19209 · Report as offensive     Reply Quote
Profile Ananas
Volunteer moderator

Send message
Joined: 31 Oct 04
Posts: 336
Credit: 3,316,482
RAC: 0
Message 19220 - Posted: 12 Jan 2006, 20:49:34 UTC - in response to Message 19209.  
Last modified: 12 Jan 2006, 20:52:09 UTC

Lord only knows why you would want to be that far out of date...


It\'s not bloated like 5.x, it\'s reliable on all my boxes, it goes through a squid with auth, I can just copy it without installing it and I have learned how to live with most of its weirdnesses :-)
ID: 19220 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 19223 - Posted: 12 Jan 2006, 21:04:55 UTC

Carl has said elsewhere that he is changing the way data is stored on computers, and also when it is sent.
This already applies to spinup, and is, apparently, going to apply to coupled ocean.
With spinup, the trickles are more frequent, and larger, than in slab/sulphur. They now contain science data, and there is also a BIG trickle at each 50 year point. These \'trickles\' need to be sent back when they are created, both so that the researchers can keep on eye on models to make sure they are stable, and so that there isn\'t a HUGE lot of upload traffic at the end.
And, at least with spinup, the researchers can send a \'kill\' message to abort an unstable model before it wastes too much computer time.

All of which needs the new 5.* version of BOINC.
The sooner you migrate to the new version, the less pain. As some classic SETI people have found out. :(

ID: 19223 · Report as offensive     Reply Quote
Profile Ananas
Volunteer moderator

Send message
Joined: 31 Oct 04
Posts: 336
Credit: 3,316,482
RAC: 0
Message 19225 - Posted: 12 Jan 2006, 22:38:19 UTC
Last modified: 12 Jan 2006, 22:43:45 UTC

As long as Berkeley keeps adding one new feature each day to keep the bugs happy, 5.x isn\'t really an option for me. Too many headless crunchers to install a bugfixed version with new surprises every week.

They should freeze features some day and start handling the problems, our team forum is stuffed with posts from people who have trouble - and in the project fora it doesn\'t look much better.
ID: 19225 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 19227 - Posted: 12 Jan 2006, 23:01:26 UTC

5.2.8 works for me. But I DO have a perfect computer, and I only run one project.
I wonder if a new minimum version will be enforced in February? Still, it won\'t bother me.

ID: 19227 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 19230 - Posted: 12 Jan 2006, 23:29:24 UTC

Ananas
Just a thought: Would it be possible to do a survey of your team members to see what they are using for these:

BOINC version
processor
os
graphics card
virus checker
firewall
proxy server
is a program being run using DirectX or 3D graphics?

main problem

Maybe there is a pattern, and someone here could pin down a cure.

ID: 19230 · Report as offensive     Reply Quote
Profile Ananas
Volunteer moderator

Send message
Joined: 31 Oct 04
Posts: 336
Credit: 3,316,482
RAC: 0
Message 19235 - Posted: 13 Jan 2006, 0:14:15 UTC
Last modified: 13 Jan 2006, 0:37:11 UTC

I did that now in this thread and in this one too.

6000+ users - not all active in the forum of course, many gave up on BOINC too - but maybe we\'ll get at least a little feedback.


p.s.: if they all reply, I\'m screwed - I can translate stuff but there are limits ;-)
ID: 19235 · Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : Server State Over, but wu is in progress!

©2024 climateprediction.net