Some time ago I updated my BOINC client to 5.4.9. Now it always says that the computer is overcommitted and won\'t run my CPDN model.

I have a dual core (AMD Athlon 64 3800+ X2) processor, and I want to run CPDN on one core and WCG on the other. In the past I have managed to do that by having CPDN allocate somewhat above 50% of the resource share, letting CPDN download a model, and then stop CPDN from downloading more job. Since the BOINC client upgrade, there is always a message that the computer is overcommitted and is using earliest deadline first. That means that it will always choose to run WCG (or malariacontrol) on both cores (it downloads two WCG/malariacontrol at the same time).

I thought at first that it was because it needed to learn how much time it takes to process a WCG workunit, but since it\'s been doing lots and lots of them by now, that seems unlikely. Could it instead by the CPDN that is causing the overcommitted message (even though the deadline is way over in 2008)?

I enclose the first lines from the BOINC manager:
2006-11-18 13:24:19||Starting BOINC client version 5.4.9 for windows_intelx86
2006-11-18 13:24:19||libcurl/7.15.3 OpenSSL/0.9.8a zlib/1.2.3
2006-11-18 13:24:19||Data directory: C:\\Program\\BOINC
2006-11-18 13:24:21|SETI@home|Found app_info.xml; using anonymous platform
2006-11-18 13:24:22||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 3800+
2006-11-18 13:24:22||Memory: 1023.23 MB physical, 3.90 GB virtual
2006-11-18 13:24:22||Disk: 76.32 GB total, 8.87 GB free
2006-11-18 13:24:22|CPDN Seasonal Attribution Project|URL: http://attribution.cpdn.org/; Computer ID: 311; location: ; project prefs: default
2006-11-18 13:24:22|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 239124; location: home; project prefs: default
2006-11-18 13:24:22|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 1497356; location: ; project prefs: default
2006-11-18 13:24:22|malariacontrol.net beta|URL: http://www.malariacontrol.net/; Computer ID: 4496; location: home; project prefs: default
2006-11-18 13:24:22|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 2536; location: Default; project prefs: default
2006-11-18 13:24:22||General prefs: from World Community Grid (last modified 2005-12-12 07:37:53)
2006-11-18 13:24:22||General prefs: no separate prefs for Default; using your defaults
2006-11-18 13:24:22||Local control only allowed
2006-11-18 13:24:22||Listening on port 31416
2006-11-18 13:24:23|climateprediction.net|Deferring task hadcm3lbm_agup_25257668_0
2006-11-18 13:24:23|World Community Grid|Resuming task faah0952_d071n043_x1AJX_00_0 using faah version 528
2006-11-18 13:24:24|World Community Grid|Resuming task faah0952_d071n044_x1AJX_00_2 using faah version 528
2006-11-18 13:24:24||Using earliest-deadline-first scheduling because computer is overcommitted.
2006-11-18 13:24:24||Suspending work fetch because computer is overcommitted.
2006-11-18 13:24:26||Suspending computation - user is active
2006-11-18 13:24:26|World Community Grid|Pausing task faah0952_d071n043_x1AJX_00_0 (left in memory)
2006-11-18 13:24:26|World Community Grid|Pausing task faah0952_d071n044_x1AJX_00_2 (left in memory)
2006-11-18 13:24:26||Suspending network activity - user is active
2006-11-18 13:25:47||Resuming computation
2006-11-18 13:25:47||Rescheduling CPU: Resuming computation
2006-11-18 13:25:47||Resuming network activity
2006-11-18 13:25:47|World Community Grid|Resuming task faah0952_d071n043_x1AJX_00_0 using faah version 528
2006-11-18 13:25:47|World Community Grid|Resuming task faah0952_d071n044_x1AJX_00_2 using faah version 528

SETI and seasonal attrib are set not to get any new workunits and therefore never runs.
If I look in the client_state file, I see that the long term debt is a large positive number for CPDN, and negative for malaria and WCG.

---

You seem to have found the problem.

BOINC is operating as it is designed to do. Unfortunately, it isn\'t always the way we want it to.

My machines run only CPDN so I have no experience on which to draw and you\'ve already tried all I know about it.

There\'s a write-up about how it works somewhere in the Wiki:
http://boinc-wiki.ath.cx/index.php?title=Main_Page

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Thanks for the answer.

BOINC is operating as it is designed to do. Unfortunately, it isn\'t always the way we want it to.

Boinc is worried that CPDN may not be finished in time. As a way of solving it, it makes sure that CPDN never runs. No, that isn\'t exactly the way I would want it to work!

Is there any number in the client_state of other file that I can change in order to make the computer believe (realize) that it isn\'t overcommited as a result of CPDN?

---

Might try setting all debt entries to zero with Notepad, after shutting down CPDN & boinc and making a backup. Format is 0.000000 for zero values.

Best of luck.

Edit: As to making boinc think it isn\'t over-committed, no way I know that would work with certainty. Possibilities in tweaking time values but I have no idea what might result from that.

Might try setting Prefs to use one CPU, though boinc has a problem with Pref updates of late. A workaround is mentioned on one of the Boards but I don\'t remember what to \'Search\' for. (I\'ll try a manual search, but am not sure it\'s on these Boards -- and the CPDN discussion Boards are still down.)

Edit2: Found a direct link to the Wiki write-up, thanks to one of Les\' posts:
http://boinc-wiki.ath.cx/index.php?title=Work_Scheduler

Edit3: Found this post; it has a link to the boinc site\'s recommended \"solution\". Hope it helps.
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=5008&nowrap=true#25054

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Thanks! I guess the easiest solution is just to let CPDN download two workunits. I\'ll give that a try!

---