Advanced search

Message boards : Number crunching : Way to exclude NOELIA tasks?

Author Message
Profile microchip
Avatar
Send message
Joined: 4 Sep 11
Posts: 110
Credit: 326,102,587
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28960 - Posted: 4 Mar 2013 | 14:32:45 UTC

Hi,

As the title says, is this possible? I have nothing but trouble with NOELIA tasks, both short and long ones, so would like to avoid crunching them as I don't want to babysit my BOINC and look for which works OK and which not.
____________

Team Belgium

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28963 - Posted: 4 Mar 2013 | 15:08:38 UTC - in response to Message 28960.

I too have been having nothing but trouble with NOELIA tasks.

On one of my machines (GTX 650 Ti) a task had run for just over 6hrs. I exited BOINC to do something else for about 30 mins, and when I started BOINC back up I immediately got a compute error task ended. After 6 hrs...wasted.

I had also seen several of these tasks error out on my other machine, so this morning I started manually aborting them as they downloaded. If possible I'm going to try and stick with TONI tasks. I don't even have this much trouble with NATHAN tasks.

Operator
____________

Profile microchip
Avatar
Send message
Joined: 4 Sep 11
Posts: 110
Credit: 326,102,587
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28965 - Posted: 4 Mar 2013 | 15:36:55 UTC

Both TONI an NATHAN tasks run without problems (short & long queues). Just NOELIA ones are very unpredictable so if someone know how to avoid them, I'd be more than happy.
____________

Team Belgium

Jorge Alberto Ramos Olive...
Send message
Joined: 13 Aug 09
Posts: 24
Credit: 156,684,745
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwat
Message 28967 - Posted: 4 Mar 2013 | 15:54:58 UTC

I think it's not possible. The only restriction I know of is of excluding short and/or long tasks.

[AF>Belgique] bill1170
Send message
Joined: 4 Jan 09
Posts: 13
Credit: 1,195,187,199
RAC: 5,433,293
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28968 - Posted: 4 Mar 2013 | 16:15:04 UTC - in response to Message 28967.
Last modified: 4 Mar 2013 | 16:15:50 UTC

I have de same problem of not having the possibility to babysit the computer.

As it crash almost once a day, unfortunately I have to crunch on another project. I'm sorry for that.I will read the forum until a solution has been found.

A good idea for the scientists could be to move the unstable tasks in the beta queue which would allow to continue to crunch the other tasks.

Profile nate
Send message
Joined: 6 Jun 11
Posts: 124
Credit: 2,928,865
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 28973 - Posted: 4 Mar 2013 | 18:07:10 UTC

http://www.gpugrid.net/forum_thread.php?id=3311&nowrap=true#28972

Operator
Send message
Joined: 15 May 11
Posts: 108
Credit: 297,176,099
RAC: 0
Level
Asn
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28996 - Posted: 6 Mar 2013 | 0:34:50 UTC

I have NOELIAs in my queue on one of my machines again, and another machine is crunching one as I write this.

Says it has taken 7:41 so far on the system in progress and estimates 17 more hrs (GTX 650 Ti SSC 2GB) for a system that normally goes through TONIs in around 7-10hrs.

I noticed on that machine that instead of using 25-27% CPU (Quad Core 9550) that it was only using 12-13% now. This is new behavior.

So are these the "Reloaded" NOELIAs?

Kind of catchy, ...NOELIA, She's reloaded, and this time it's personal!!

I'm still afraid to pause, or exit from BOINC to do other work in case it errors out with a computational error, especially since it's taken almost 8hrs so far!

That would be a waste.

Operator


____________

Profile microchip
Avatar
Send message
Joined: 4 Sep 11
Posts: 110
Credit: 326,102,587
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 28999 - Posted: 6 Mar 2013 | 10:53:05 UTC

Well, NOELIA submitted new WUs on the short 4.2 queue and I've crunched 3 so far and had no problems. They usually crunch in 2.5 hours here and credit is 1500 points. I noticed, though, that they don't fully stress the GPU but I'm OK with that
____________

Team Belgium

Profile Steve Hawker*
Send message
Joined: 12 Feb 13
Posts: 5
Credit: 1,926,775
RAC: 0
Level
Ala
Scientific publications
watwatwat
Message 29063 - Posted: 7 Mar 2013 | 21:53:28 UTC - in response to Message 28963.
Last modified: 7 Mar 2013 | 21:55:43 UTC

On one of my machines (GTX 650 Ti) a task had run for just over 6hrs. I exited BOINC to do something else for about 30 mins, and when I started BOINC back up I immediately got a compute error task ended. After 6 hrs...wasted.

Operator


I had exactly this problem on a Long Task. Except that I was 56 hours into that task.

In the FAQ it says:

As a general rule, workunits are NEVER canceled when they are running so that you don't lose any credits or compute power.

WUs may be canceled only while they are sitting in your queue, waiting to be executed. This is normal, and in fact it is a feature of the BOINC scheduler which prevents "redundant results" to be computed.


Well, I lost 56,000 points and 56 hours of compute power.

Will the admins put in place a scheme to pay us those lost credits? Would sure be appreciated.

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 817,865,789
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29073 - Posted: 8 Mar 2013 | 7:37:27 UTC

Hum really?? I lost everytime the credits when workunits got canceld o.O
____________
DSKAG Austria Research Team: http://www.research.dskag.at



ExtraTerrestrial Apes
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 17 Aug 08
Posts: 2705
Credit: 1,311,122,549
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 29084 - Posted: 8 Mar 2013 | 21:31:27 UTC

Considering that these tasks seem to be able to BSOD a machine and/or fail and afterwards leave them in a state so that subsequent WUs are stuck.. aborting the tasks right now may have saved you (and others) even more trouble. The choice between pest and cholera is always tough.

MrS
____________
Scanning for our furry friends since Jan 2002

Post to thread

Message boards : Number crunching : Way to exclude NOELIA tasks?

//