Advanced search

Forums : Technical Support : Longer / heavier WUs?
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · Next

AuthorMessage
Profile sysfried

Send message
Joined: 24 Jun 07
Posts: 114
Credit: 5,296,905
RAC: 0
Message 8104 - Posted: 15 Apr 2009, 11:26:43 UTC - in response to Message 8100.  



In the mean time, these tasks can be run to completion (if you have the memory available).


I love to support Cosmo@Home and I have the RAM avail, but sorry, nope, not going to happen. Please clean out the available workunits.

Thank you very much.

Cheers,

Sysfried.
Happy member of Team: Planet 3D Now!

ID: 8104 · Report as offensive     Reply Quote
Matthias Lehmkuhl

Send message
Joined: 29 Aug 07
Posts: 15
Credit: 1,013,277
RAC: 235
Message 8108 - Posted: 16 Apr 2009, 8:09:20 UTC

I got also some of this large results - all Win 32
resultid=12949759
CPU time (sec): 70,915.81
claimed credit: 291.25
granted credit: 140.00

Checkpointing was round 2 hours or more.
Memory was also round 350MB + VM

resultid=12937148
CPU time (sec): 32,862.02

one more is not finished
resultid=13050156
7 hours runtime at 70% now
actual memory 761 MB + VM 760 MB
Matthias
ID: 8108 · Report as offensive     Reply Quote
Tomas

Send message
Joined: 18 Aug 08
Posts: 30
Credit: 827,240
RAC: 0
Message 8146 - Posted: 19 Apr 2009, 15:24:56 UTC

The new WUs are still using a lot of memory, 300 - 500 MB.
Is this supposed to be normal?
ID: 8146 · Report as offensive     Reply Quote
zpm

Send message
Joined: 13 Mar 09
Posts: 3
Credit: 8,400
RAC: 0
Message 8149 - Posted: 19 Apr 2009, 21:02:34 UTC - in response to Message 8108.  

I got also some of this large results - all Win 32
resultid=12949759
CPU time (sec): 70,915.81
claimed credit: 291.25
granted credit: 140.00

Checkpointing was round 2 hours or more.
Memory was also round 350MB + VM

resultid=12937148
CPU time (sec): 32,862.02

one more is not finished
resultid=13050156
7 hours runtime at 70% now
actual memory 761 MB + VM 760 MB



1 of mine has 5 hours under it's belt and only 40%(etc 6+ hrs.)... a couple others that started later than that one, have caught up to that one.
ID: 8149 · Report as offensive     Reply Quote
Tomas

Send message
Joined: 18 Aug 08
Posts: 30
Credit: 827,240
RAC: 0
Message 8158 - Posted: 20 Apr 2009, 9:03:52 UTC

I aborted one of the 2 WU I downloaded, in total they used 1.2GB ram
and 50% of all disk read\write operation related to BOINC.






ID: 8158 · Report as offensive     Reply Quote
Ryo

Send message
Joined: 21 Apr 09
Posts: 1
Credit: 5,500
RAC: 0
Message 8197 - Posted: 22 Apr 2009, 12:03:06 UTC
Last modified: 22 Apr 2009, 12:54:33 UTC

I have 3 units and those using about 1,1GB.
I'm sorry, but this is way, way too much. If this isn't a bug, or a memory leak, I would please to reduce the amount of mem needed dramatically.

My computer isn't completely usable anymore, and had to stop work, because I got "out of memory" errors.

See screenshot:
-> http://drop.io/hidden/jciafrrlh2ofll/asset/Y29zbW9sb2d5YXRob21lLW1lbS1qcGc=
ID: 8197 · Report as offensive     Reply Quote
Gary Wilson

Send message
Joined: 30 Nov 08
Posts: 8
Credit: 5,000,011
RAC: 0
Message 8204 - Posted: 23 Apr 2009, 1:31:17 UTC - in response to Message 8197.  

I reattached to the project after the 2.16 update was posted, but no good. Once one of the projects starting running, the other two projects running on the same box (mindmodeling, intelligence realm) started bailing out of all their WUs with computation errors. I saw this before with the busted 2.15 version and detached. I detached again and will wait for awhile I guess. I'm running 32-bit Ubuntu 8.10 with 6.4.5 on a dual-core Intel. Not sure if any other setup has seen this where multiple projects are running.
ID: 8204 · Report as offensive     Reply Quote
Profile Misfit
Volunteer tester
Avatar

Send message
Joined: 9 Jun 07
Posts: 150
Credit: 237,789
RAC: 0
Message 8205 - Posted: 23 Apr 2009, 1:45:40 UTC - in response to Message 8204.  

Who knows, you may get boatloads of free credits for those errored WUs.
me@rescam.org
ID: 8205 · Report as offensive     Reply Quote
Profile Ageless
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 15 Jun 07
Posts: 345
Credit: 50,500
RAC: 0
Message 8215 - Posted: 24 Apr 2009, 17:16:51 UTC - in response to Message 8197.  
Last modified: 24 Apr 2009, 17:17:43 UTC

I sent Anshul a couple of screen shots from my computer, showing one task taking up 1.24GB of memory (647MB RAM + VM + 'normal Windows stuff' loaded). His answer to me:

It looks to me like WUs for 2.16 have high precision floating point calculations. I will speak to Scott since he released this new version a couple of weeks ago. I may be able to refine the code enough to reduce the memory usage but I cannot guarantee anything. If this keeps happening we will split the work load for each workunit in 2 and maybe run "Half" universe simulations per WU.
I'll keep you posted.

Anshul Kanakia

Jord.
ID: 8215 · Report as offensive     Reply Quote
glaesum

Send message
Joined: 30 Apr 08
Posts: 2
Credit: 16,800
RAC: 0
Message 8217 - Posted: 24 Apr 2009, 23:37:05 UTC

Camb used to run quite happily on my ancient boat-anchor since this time last year, but now it don't like it at all... :-(
I now get this red message:
Message from server: No work sent
Message from Server: CAMB needs 486.84MB RAM, only 383.07MB is available for use.

looks like I'll have to bow out now if resource requirement stays at this level.
(obviously I'm not going to add ram to a m'ch so old)
ID: 8217 · Report as offensive     Reply Quote
Profile Misfit
Volunteer tester
Avatar

Send message
Joined: 9 Jun 07
Posts: 150
Credit: 237,789
RAC: 0
Message 8218 - Posted: 25 Apr 2009, 0:08:32 UTC

With the great increase in the amount of crunch time I've had to detach the older AMD from the project. It just couldn't finish by deadline.
me@rescam.org
ID: 8218 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 8224 - Posted: 25 Apr 2009, 18:04:19 UTC - in response to Message 8218.  
Last modified: 25 Apr 2009, 18:11:05 UTC

With the great increase in the amount of crunch time I've had to detach the older AMD from the project. It just couldn't finish by deadline.


What "older AMD"? You're showing an X2 4200+ that you don't seem to have that heavily loaded with other projects, unless it is working on a SETI AP task...


I went ahead and grabbed a few tasks from here. Looks like the first one I got, I got lucky with, as it completed in just 8.3 hours. The one I have now should complete between 15.5 and 16 hours. Memory usage is currently 804,488K @ 93% complete.

Anyway, the question comes to my mind as to why one would release a changed application version in the middle of recovering from a server crash and not even mention the potential for this large of a difference in resource usage and processing times... Sure, the post on the home page says that there is an increase, but it doesn't indicate the potential for double or triple the resource requirements / processing times.

Of course, I'm just a "continual complainer"...so what do I know?
ID: 8224 · Report as offensive     Reply Quote
Profile Misfit
Volunteer tester
Avatar

Send message
Joined: 9 Jun 07
Posts: 150
Credit: 237,789
RAC: 0
Message 8226 - Posted: 25 Apr 2009, 20:49:09 UTC - in response to Message 8224.  

With the great increase in the amount of crunch time I've had to detach the older AMD from the project. It just couldn't finish by deadline.

What "older AMD"? You're showing an X2 4200+ that you don't seem to have that heavily loaded with other projects, unless it is working on a SETI AP task...

Did I make that up or are you assuming that computer runs 24/7?
me@rescam.org
ID: 8226 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 8227 - Posted: 25 Apr 2009, 21:09:42 UTC - in response to Message 8226.  

With the great increase in the amount of crunch time I've had to detach the older AMD from the project. It just couldn't finish by deadline.

What "older AMD"? You're showing an X2 4200+ that you don't seem to have that heavily loaded with other projects, unless it is working on a SETI AP task...

Did I make that up or are you assuming that computer runs 24/7?


It's just that I'd define "older AMD" that isn't able to make the deadline as something in the AthlonXP era...unless it's overextended. However, something that might be making a difference is the 4200+ only has 512K cache per core...
ID: 8227 · Report as offensive     Reply Quote
Erik Jan Meijer
Avatar

Send message
Joined: 25 Oct 07
Posts: 1
Credit: 99,791
RAC: 0
Message 8228 - Posted: 26 Apr 2009, 8:33:44 UTC

To me it seems C@H with the new WU's now is only suited for PCs one, two years of age at most. The deadline is way too tight for my six year old PC. It was happily running older WU's in its spare time (not 24/7). But now it can't finish anything in time anymore.
It's currently working on two WU's sent to me on April 19th, their deadlines set at April 29 and the progress is still only on 39%. Remaining time is utterly wrong and only increases in time.
So I won't be processing any C@H WU's anymore...
ID: 8228 · Report as offensive     Reply Quote
Kenny Frew

Send message
Joined: 15 Apr 09
Posts: 1
Credit: 161,631
RAC: 0
Message 8230 - Posted: 26 Apr 2009, 18:21:46 UTC
Last modified: 26 Apr 2009, 18:27:59 UTC

I noticed that at 91% to 96% it repeats and starts over at 92% - 3 times now-29th getting closer! Thirty one hours so far. I have 4 gb's memory and see a drastic lag in other computor functions.
ID: 8230 · Report as offensive     Reply Quote
Gary Wilson

Send message
Joined: 30 Nov 08
Posts: 8
Credit: 5,000,011
RAC: 0
Message 8297 - Posted: 2 May 2009, 21:10:56 UTC - in response to Message 8230.  

Does anyone know if the work units have returned to the way they were before the server crash, or do they still take too many system resources to crunch.
Thanks.
ID: 8297 · Report as offensive     Reply Quote
John Price

Send message
Joined: 2 Mar 08
Posts: 4
Credit: 734,270
RAC: 0
Message 8298 - Posted: 2 May 2009, 22:51:53 UTC

I have a problem with new Cosmology@home WU. All seems to have changed since the Server Crash a fee weeks ago. It is a Linux 10 AMD 2400+ with 1Gb of RAM. It was dedicated to Cosmo, but as I was having problems I reset the project. When restarted it dowloaded 3 units the first gets to about 60% and then it says "Waiting for Memory" up to limit 935 Mb. It then starts the second unit which gets to about the 60% mark and then stalls with the same message. Then the third starts with the same problem 3-4 days or so now with no results sent.
This is all a bit pointless so I am now moving back to Climateprediction which is always able to be run.
I am still producing Cosmos on more powerfull machines and will keep an eye on this forum to see if anything changes.
I really think the work units should be able to be run on an AMD 2400+ with 1Gb of ram.
ID: 8298 · Report as offensive     Reply Quote
Guido Platteau

Send message
Joined: 11 Apr 08
Posts: 1
Credit: 37,800
RAC: 0
Message 8355 - Posted: 19 May 2009, 8:39:18 UTC

My PC can't crunch Cosmology anymore. I get the message:
19/05/2009 10:31:54 Cosmology@Home Message from server: CAMB needs 476.84 MB RAM but only 460.33 MB is available for use.
ID: 8355 · Report as offensive     Reply Quote
Bill & Patsy
Avatar

Send message
Joined: 27 Jul 08
Posts: 25
Credit: 1,045,640
RAC: 0
Message 8363 - Posted: 23 May 2009, 21:25:10 UTC - in response to Message 8355.  

My PC can't crunch Cosmology anymore. I get the message:
19/05/2009 10:31:54 Cosmology@Home Message from server: CAMB needs 476.84 MB RAM but only 460.33 MB is available for use.

Increase your Virtual Memory (Paging File size).
--Bill

ID: 8363 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · Next

Forums : Technical Support : Longer / heavier WUs?