Welcome to MilkyWay@home

No Work Given Until Run Out of Work?

Message boards : Number crunching : No Work Given Until Run Out of Work?
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 14331 - Posted: 8 Mar 2009, 9:40:24 UTC
Last modified: 8 Mar 2009, 9:55:17 UTC

With the GPU app I'm seeing the behaviour atm where while I've got less than the 6 wu's per core cached, I can hit update and BOINC (6.6.12 and XP) says it is asking for work, but MW gives me none. I can then run out of work, hit update and I get work, but not always, but currently I never get new work while there is at least 1 wu cached.

Frustrating when you've got a beast to feed....and I hate micro-managing BOINC.

Any ideas on how to overcome this?

Also, any ideas on how to get the back-off time shown again?
ID: 14331 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GalaxyIce
Avatar

Send message
Joined: 6 Apr 08
Posts: 2018
Credit: 100,142,856
RAC: 0
Message 14332 - Posted: 8 Mar 2009, 9:54:05 UTC - in response to Message 14331.  

With the GPU app I'm seeing the behaviour atm where while I've got less than the 6 wu's per core cached, I can hit update and BOINC (6.6.12 and XP) says it is asking for work, but MW gives me none. I can then run out of work, hit update and I get work, but not always, but currently I never get new work while there is at least 1 wu cached.

Frustrating when you've got a beast to feed....and I hate micro-managing BOINC.

Any ideas on how to overcome this?

Also, any ideas on how to get the back-off time shown again?

Also,

Strange things can happen when you first kick off a ATI GPU app. I've installed and re-installed the 4850 as part of running some tests, and similar to what you describe happens. Just keeping pressing update when it gets 'stuck' and eventually it will settle down to a pattern where you can just watch it ripple without any micro-managing.

If it doesn't settle down... then also,


ID: 14332 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Conan
Avatar

Send message
Joined: 2 Jan 08
Posts: 122
Credit: 69,479,692
RAC: 1,456
Message 14337 - Posted: 8 Mar 2009, 11:13:49 UTC - in response to Message 14331.  

G'Day The Gas Giant,
I wouldn't fret too much about this as I have the same problem with a lowly Pentium P4 253, single processor.
It will wait till the 6th and last work unit is complete before it downloads any new work.
Hitting update, even if the last WU is nearly complete just gets you a "Requesting 0 seconds of new work, 0 work sent" type of message.
This has been going on for a while, it also affects a couple of my Linux machines as well but I think in their case it is due to BM owing time to other projects.

Conan.

ID: 14337 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
val73bg

Send message
Joined: 5 Oct 08
Posts: 5
Credit: 15,638,470
RAC: 0
Message 14339 - Posted: 8 Mar 2009, 11:24:11 UTC

@The Gas Giant : What is your "time to completion" in Boinc manager? Is it real? If there is high values ,no work will download, until last workunit.
ID: 14339 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 14340 - Posted: 8 Mar 2009, 11:49:39 UTC

Time completion is a very real 30 seconds...but I changed the resource share for SETI (reduced it) so that is most likely the cause. Maybe I'll abort a few of SETI wu's and see if that helps.
ID: 14340 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
val73bg

Send message
Joined: 5 Oct 08
Posts: 5
Credit: 15,638,470
RAC: 0
Message 14342 - Posted: 8 Mar 2009, 12:01:53 UTC

I'm running 4 projects.Two of them active, the other two suspended. The resource share is 100% for each(by default). In Boinc manager shown 100%(25.00%) for each one.No problems with sharing quota for now :).
ID: 14342 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Bill & Patsy
Avatar

Send message
Joined: 7 Jul 08
Posts: 47
Credit: 13,629,944
RAC: 0
Message 14414 - Posted: 8 Mar 2009, 19:47:02 UTC

Yes. It could be a resource share issue when you're running more than one project. Long Term Debt heavily influences how many work units will download. BOINC Debt Viewer is very nice for watching this.
--Bill

ID: 14414 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 14426 - Posted: 8 Mar 2009, 20:54:39 UTC
Last modified: 8 Mar 2009, 20:55:00 UTC

Yes, but I can hit update, BOINC says it is requesting new tasks but 0 new tasks are received until I completely run out of MW work.
ID: 14426 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 14427 - Posted: 8 Mar 2009, 20:55:47 UTC - in response to Message 14426.  

Yes, but I can hit update, BOINC says it is requesting new tasks but 0 new tasks are received until I completely run out of MW work.


That's generally how it's always been for me. I'll be down to 1 task and want more so I don't run out and can't get any.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 14427 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 14510 - Posted: 9 Mar 2009, 11:28:21 UTC

<return_results_immediately>1</return_results_immediately>

Does the trick. Still don't get any wu's while my comp has work, but atleast my comp gets some once it returns the last wu immediately. Hate to load the server unnecessarily, but.....
ID: 14510 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Phil
Avatar

Send message
Joined: 13 Feb 08
Posts: 1124
Credit: 46,740
RAC: 0
Message 14531 - Posted: 9 Mar 2009, 15:14:52 UTC - in response to Message 14510.  

<return_results_immediately>1</return_results_immediately>


Yes adding that option to cc_config chivvies it along a little.
ID: 14531 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 14544 - Posted: 9 Mar 2009, 17:08:00 UTC

Just noticed my comp had no MW work and has been like that for the last 5 hrs with a back off of 19hrs left. !@#$&^. Hitting update got it 24 tasks.

Is there some sort of speed limit being implemented by Travis?
ID: 14544 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 14545 - Posted: 9 Mar 2009, 17:16:46 UTC - in response to Message 14544.  

Just noticed my comp had no MW work and has been like that for the last 5 hrs with a back off of 19hrs left. !@#$&^. Hitting update got it 24 tasks.

Is there some sort of speed limit being implemented by Travis?


Probably go 'no work' a few times and it backed the project off. The only limit is 5000 wu's/day. Some have had trouble getting work recently.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 14545 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Misfit
Avatar

Send message
Joined: 27 Aug 07
Posts: 915
Credit: 1,503,319
RAC: 0
Message 14548 - Posted: 9 Mar 2009, 17:27:55 UTC - in response to Message 14531.  

<return_results_immediately>1</return_results_immediately>

Yes adding that option to cc_config chivvies it along a little.

Sometimes I'll still have to dump a few Cosmo WUs before MW will pick up anything.
me@rescam.org
ID: 14548 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : No Work Given Until Run Out of Work?

©2024 Astroinformatics Group