"reached a limit on tasks in progress"


log in

Advanced search

Message boards : Number crunching : "reached a limit on tasks in progress"

Previous · 1 · 2
Author Message
Greg W Jones
Send message
Joined: 28 Nov 02
Posts: 29
Credit: 25,473,033
RAC: 120
United States
Message 1183889 - Posted: 8 Jan 2012, 22:38:48 UTC - in response to Message 1182325.

BAPRIEBE

I was having a similar issue of not getting seti WUs when I had several other boinc projects suscribed on the same computer. As others have already stated, boinc will fill up your cache with the other projects when it cannot get WUs from seti. This issue is compounded with the very long retry delays in boinc 6.12.34, that you are running.

So, based on recommendations from other posters here (Thank you, Claggy), I switched to boinc 6.10.60. Now, I typically have a full cache near the WU limits.

What I have noticed with boinc 6.10.60 is that its WU download retry intervals are short. Therefore, I don't get into a situation where an unattended boinc rig has a 4 hour delay before trying to download a WU again. This extended delay allows the cache to be filled by the other projects.

BTW, I think the seti WU limits are 50 per physical CPU core and 300 per GPU.

Cheers,

____________

Profile SciManStev
Volunteer tester
Avatar
Send message
Joined: 20 Jun 99
Posts: 4792
Credit: 79,796,670
RAC: 35,805
United States
Message 1183892 - Posted: 8 Jan 2012, 22:59:08 UTC

50 per CPU core, and 400 per GPU are the current limits. Double CPU's help, such as GTX 295's, as they double the amount of work allowed.

Steve
____________
Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

Profile arkayn
Volunteer tester
Avatar
Send message
Joined: 14 May 99
Posts: 3595
Credit: 47,380,702
RAC: 4,180
United States
Message 1183898 - Posted: 8 Jan 2012, 23:08:32 UTC - in response to Message 1183892.

50 per CPU core, and 400 per GPU are the current limits. Double CPU's help, such as GTX 295's, as they double the amount of work allowed.

Steve


Yep, I have 195+4 CPU and 796+2 GPU on board currently.
____________

Dave
Avatar
Send message
Joined: 29 Mar 02
Posts: 774
Credit: 23,193,139
RAC: 0
United Kingdom
Message 1184086 - Posted: 9 Jan 2012, 17:32:27 UTC - in response to Message 1183892.

50 per CPU core, and 400 per GPU are the current limits. Double CPU's help, such as GTX 295's, as they double the amount of work allowed.

Steve


Does it detect HT on/off or app_info changes? Is that actual physical core or thread?

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4048
Credit: 32,693,420
RAC: 508
United Kingdom
Message 1184090 - Posted: 9 Jan 2012, 17:42:07 UTC - in response to Message 1184086.
Last modified: 9 Jan 2012, 17:49:32 UTC

50 per CPU core, and 400 per GPU are the current limits. Double CPU's help, such as GTX 295's, as they double the amount of work allowed.

Steve


Does it detect HT on/off or app_info changes? Is that actual physical core or thread?

That's total number of threads, and GPUs, my i7-2600K/GTX460/HD5770 has a limit of 1200, 400 per GPU, and 400 for the CPU (8*50),

Claggy

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,196,591
RAC: 100,183
United States
Message 1184111 - Posted: 9 Jan 2012, 18:34:30 UTC - in response to Message 1184086.

50 per CPU core, and 400 per GPU are the current limits. Double CPU's help, such as GTX 295's, as they double the amount of work allowed.

Steve


Does it detect HT on/off or app_info changes? Is that actual physical core or thread?

If you don't modify anything it will use the total # of cores, virtual and real. If you have imposed a limit via "use at most x% of CPUs" or have limited the processors in your cc_config.xml then that will be used.

If is you set use at most 75% of CPUs, in a 8 core system, and have BOINC to not use 1 of 2 GPUs for S@H. Then you would get 300 CPU and 400 GPU tasks instead of 400 & 800.

____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile Khangollo
Avatar
Send message
Joined: 1 Aug 00
Posts: 245
Credit: 36,410,524
RAC: 0
Slovenia
Message 1184131 - Posted: 9 Jan 2012, 19:24:00 UTC

There's also the <ncpus> option in cc_config.xml to override boinc's CPU count detection. It can be used in combination with "use at most % of cores" option in preferences...

...
____________

hbomber
Volunteer tester
Send message
Joined: 2 May 01
Posts: 437
Credit: 50,852,768
RAC: 216
Bulgaria
Message 1184321 - Posted: 10 Jan 2012, 12:20:46 UTC

I have overridden number of cores by modifying cc_config, from 4 to 2, and I still get 200 CPU workunits. My limit is 1000 atm and I have two GPUs, while, as written, only 2 cores are acive.
____________

Profile Firehawk
Volunteer tester
Avatar
Send message
Joined: 21 May 99
Posts: 1708
Credit: 252,501,428
RAC: 72,477
Brazil
Message 1184444 - Posted: 11 Jan 2012, 2:43:07 UTC

I keep having this message even with zero GPU tasks and like 40 cpu tasks in a six core/five gpu machine for weeks now.
It just happens with my biggest rig, the others, smallers, get much much more work. Frustrating to say the least. BTW I guess this is the case with all top crunchers, they are all out of work and not getting any.
____________

Josef W. Segur
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4206
Credit: 1,030,844
RAC: 270
United States
Message 1184595 - Posted: 11 Jan 2012, 17:52:10 UTC - in response to Message 1184444.

I keep having this message even with zero GPU tasks and like 40 cpu tasks in a six core/five gpu machine for weeks now.

That indicates some problem, obviously. Since your computers are hidden we can't get any hints from the web pages.

It just happens with my biggest rig, the others, smallers, get much much more work. Frustrating to say the least. BTW I guess this is the case with all top crunchers, they are all out of work and not getting any.

Checking the first ten top hosts, on average they've gotten 759 GPU (MB) tasks today (Berkeley time, approaching 10 am now).
Joe

Profile RottenMutt
Avatar
Send message
Joined: 15 Mar 01
Posts: 992
Credit: 207,654,623
RAC: 1
United States
Message 1184749 - Posted: 12 Jan 2012, 13:50:00 UTC

something is all hosed up again on the server side as my small queues are draining again this morning...
____________

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3868
Credit: 107,196,591
RAC: 100,183
United States
Message 1184751 - Posted: 12 Jan 2012, 14:05:34 UTC - in response to Message 1184749.

something is all hosed up again on the server side as my small queues are draining again this morning...

I'm sorry it must have been my machine sucking up all of the work. Looking at the logs from overnight for my machines most of them are bouncing off of their limits & I don't see any failed connections to the scheduler. Also I only seem to have 13 transfers across the ~30 machines.

Do you have a lot of active transfers? Which would cause BOINC not to request new tasks. I remember that n cups * 2 is one of the formulas used to stop requesting work, but I think that is for uploads.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile Wiggo
Avatar
Send message
Joined: 24 Jan 00
Posts: 6532
Credit: 90,649,816
RAC: 75,256
Australia
Message 1184798 - Posted: 12 Jan 2012, 19:23:21 UTC - in response to Message 1184751.

Partly my fault as well seeing as all mine have been feeding very well but still bouncing off/on their limits.

Cheers.
____________

Profile Michel448a
Volunteer tester
Avatar
Send message
Joined: 27 Oct 00
Posts: 1201
Credit: 2,891,635
RAC: 0
Canada
Message 1184857 - Posted: 13 Jan 2012, 0:25:59 UTC - in response to Message 1184751.
Last modified: 13 Jan 2012, 1:20:00 UTC


I'm sorry it must have been my machine sucking up all of the work. Looking at the logs from overnight for my machines most of them are bouncing off of their limits & I don't see any failed connections to the scheduler. Also I only seem to have 13 transfers across the ~30 machines.


HAHA

i guess ..
i shall be happy with my poor little machine
he is working as hard as...
a little mountain dwarf :)

Profile Firehawk
Volunteer tester
Avatar
Send message
Joined: 21 May 99
Posts: 1708
Credit: 252,501,428
RAC: 72,477
Brazil
Message 1184867 - Posted: 13 Jan 2012, 0:57:03 UTC - in response to Message 1184595.
Last modified: 13 Jan 2012, 1:01:19 UTC

I keep having this message even with zero GPU tasks and like 40 cpu tasks in a six core/five gpu machine for weeks now.

That indicates some problem, obviously. Since your computers are hidden we can't get any hints from the web pages.

It just happens with my biggest rig, the others, smallers, get much much more work. Frustrating to say the least. BTW I guess this is the case with all top crunchers, they are all out of work and not getting any.

Checking the first ten top hosts, on average they've gotten 759 GPU (MB) tasks today (Berkeley time, approaching 10 am now).
Joe


They aren´t hidden anymore. Even then, my top rig is at #1 position right now, so i´m pretty sure that´s there is a problem, but it´s not with my machine, because I don´t own the most powerfull machine out there. So the big guns aren´t having work on the past days, obviously. Have to say that that machine it´s having work on the past 24 hours, without any change on my side too, and even building a cache, so something was fixed at that time.
____________

Profile Firehawk
Volunteer tester
Avatar
Send message
Joined: 21 May 99
Posts: 1708
Credit: 252,501,428
RAC: 72,477
Brazil
Message 1184932 - Posted: 13 Jan 2012, 11:35:58 UTC

Forget the "cache" thing, they are all running dry again.
____________

Previous · 1 · 2

Message boards : Number crunching : "reached a limit on tasks in progress"

Copyright © 2014 University of California