It is running as the only project on a P4 Windows computer. The task manager shows 100% of the CPU capacity taken regardless of what I set the preferences for. I want it to suspend when computer is active, but that does not seem to work.

I have included the message list. I manually \"snooze\" the project when I am using the computer, but that is a pain, and I notice that it will resume after about 30 minutes of snoozing. It must be well rested after that period! ;-)

Thomas

1/19/2007 17:47:51||Starting BOINC client version 5.4.11 for windows_intelx86
1/19/2007 17:47:51||libcurl/7.15.3 OpenSSL/0.9.8a zlib/1.2.3
1/19/2007 17:47:51||Data directory: C:\\Program Files\\BOINC
1/19/2007 17:47:51||Processor: 2 GenuineIntel Intel(R) Pentium(R) 4 CPU 3.20GHz
1/19/2007 17:47:51||Memory: 511.29 MB physical, 1.22 GB virtual
1/19/2007 17:47:51||Disk: 143.52 GB total, 70.23 GB free
1/19/2007 17:47:52|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 511032; location: ; project prefs: default
1/19/2007 17:47:52||No general preferences found - using BOINC defaults
1/19/2007 17:47:52||Local control only allowed
1/19/2007 17:47:52||Listening on port 31416
1/19/2007 17:47:52||Suspending network activity - running CPU benchmarks
1/19/2007 17:47:54||Running CPU benchmarks
1/19/2007 17:48:53||Benchmark results:
1/19/2007 17:48:53|| Number of CPUs: 2
1/19/2007 17:48:53|| 1420 floating point MIPS (Whetstone) per CPU
1/19/2007 17:48:53|| 1286 integer MIPS (Dhrystone) per CPU
1/19/2007 17:48:53||Finished CPU benchmarks
1/19/2007 17:48:54|climateprediction.net|Resuming task hadcm3ohe_1jxa_05720816_0 using hadcm3 version 515
1/19/2007 17:48:55|climateprediction.net|Resuming task hadcm3ohe_1flf_05715205_0 using hadcm3 version 515
1/19/2007 17:48:55||Resuming computation
1/19/2007 17:48:55||Rescheduling CPU: Resuming computation
1/19/2007 17:48:55||Resuming network activity
1/19/2007 17:59:49||Suspending computation - user request
1/19/2007 17:59:49|climateprediction.net|Pausing task hadcm3ohe_1jxa_05720816_0 (removed from memory)
1/19/2007 17:59:49|climateprediction.net|Pausing task hadcm3ohe_1flf_05715205_0 (removed from memory)
1/19/2007 17:59:49||Suspending network activity - user request
1/19/2007 18:59:49||Resuming computation
1/19/2007 18:59:49||Rescheduling CPU: Resuming computation
1/19/2007 18:59:49||Resuming network activity
1/19/2007 18:59:49|climateprediction.net|Restarting task hadcm3ohe_1jxa_05720816_0 using hadcm3 version 515
1/19/2007 18:59:49|climateprediction.net|Restarting task hadcm3ohe_1flf_05715205_0 using hadcm3 version 515
1/19/2007 20:15:04||Suspending computation - user request
1/19/2007 20:15:04|climateprediction.net|Pausing task hadcm3ohe_1jxa_05720816_0 (removed from memory)
1/19/2007 20:15:04|climateprediction.net|Pausing task hadcm3ohe_1flf_05715205_0 (removed from memory)
1/19/2007 20:15:04||Suspending network activity - user request
1/19/2007 20:19:53||Resuming computation
1/19/2007 20:19:53||Rescheduling CPU: Resuming computation
1/19/2007 20:19:53||Resuming network activity
1/19/2007 20:19:53|climateprediction.net|Restarting task hadcm3ohe_1jxa_05720816_0 using hadcm3 version 515
1/19/2007 20:19:53|climateprediction.net|Restarting task hadcm3ohe_1flf_05715205_0 using hadcm3 version 515
1/20/2007 07:10:43||Suspending computation - user request
1/20/2007 07:10:43|climateprediction.net|Pausing task hadcm3ohe_1jxa_05720816_0 (removed from memory)
1/20/2007 07:10:43|climateprediction.net|Pausing task hadcm3ohe_1flf_05715205_0 (removed from memory)
1/20/2007 07:10:43||Suspending network activity - user request
1/20/2007 08:10:42||Resuming computation
1/20/2007 08:10:42||Rescheduling CPU: Resuming computation
1/20/2007 08:10:42||Resuming network activity
1/20/2007 08:10:43|climateprediction.net|Restarting task hadcm3ohe_1jxa_05720816_0 using hadcm3 version 515
1/20/2007 08:10:43|climateprediction.net|Restarting task hadcm3ohe_1flf_05715205_0 using hadcm3 version 515
1/20/2007 08:46:51||Suspending computation - user request
1/20/2007 08:46:51|climateprediction.net|Pausing task hadcm3ohe_1jxa_05720816_0 (removed from memory)
1/20/2007 08:46:51|climateprediction.net|Pausing task hadcm3ohe_1flf_05715205_0 (removed from memory)
1/20/2007 08:46:51||Suspending network activity - user request

By design, boinc projects use everything you don\'t; that\'s why you see 100% across the board.

There is a problem with loading changes made to Prefs. As far as I know, this is changed in the 5.8.x versions.

Same is also true for the capability to reduce total CPU usage. Note, however, that CPU reduction is a clock-on/clock-off process, not a throttle-back, i.e., if you set 50%, it\'ll run at 100% for awhile, then idle for similar period.

Edit: Your progress will be greater if you set leave-in-memory to Yes.

There is a way to create a file to modify your Prefs locally; instructions are on the boinc site. Sorry, I don\'t have the link.

If you set \"No new work\" in the Projects tab, then suspend one of the Models, Task Manager will report 50% -- but actually be using about 80-85% of your HT P4.

---

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

If cpdn preference changes are still not taking effect, another way of updating them might be to attach to another boinc project and change them within that project. The change should then propagate to cpdn. You could then select No new work for that project using the button in the boinc manager.

But don\'t do it using SETI as I think they\'re experiencing the same problem with their server.

---

Cpdn news

I also have a problem where I set ClimatePrediction to run for 120 minutes. I left it overnight and it went for 11 hours straight. I have 4 other tasks, why won\'t they switch out?

Hi Travis, welcome to the forum.

Just so we\'re sure what we\'re talking about, ClimatePrediction (cpdn) is a project and each model is a task or workunit.

You appear to have 4, not 5, workunits:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=542124

Your computer is dual-core so the most you should have is one model for each core. You probably clicked Suspend in the Tasks tab. This suspends the workunit (=task) but not the project, so boinc finds itself idle and promptly fetches more work to do.

I suggest that for each core, you select the model that\'s done least work, highlight it and click Abort. You could click the No new work button as well.

I suggest a look at the 4 READMEs in the News and Announcements section here:

http://www.climateprediction.net/board/index.php

The Running the model selection is very useful, plus items 1 and 5 in the README about avoiding model crashes.

---

Cpdn news

When I said 5 tasks, 2 were from Climateprediction, 2 from CPDN Seasonal Attribution, one from Einstine, and one from World Community Grid.
When I run CPU bookmarks, it reports 1 CPU, and when I bought my computer, it said it was a Hyperthreading Penteam 4, and I know that a Hyperthreading processer will report as a duel core processor.

I am not sure if it is switching out now or not, I have been doing it manually. Einstein@Home and World Community do switch out, so right now I am letting those two finnish and not get any new tasks, then I will work them only on my Mac.

Is HyperThreading shut off in the BIOS?

Edit: One sometimes sees \"strange\" run-behavior triggered by the way boinc manages resources. It tries to get everything done before deadline and if that means stealing time-slices from another Project, so be it. In so doing, it keeps track and builds-up long-term debt, to be repaid when opportunity presents itself.

---

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

Is HyperThreading shut off in the BIOS?

Edit: One sometimes sees \"strange\" run-behavior triggered by the way boinc manages resources. It tries to get everything done before deadline and if that means stealing time-slices from another Project, so be it. In so doing, it keeps track and builds-up long-term debt, to be repaid when opportunity presents itself.

HyperThreading is indeed on. I guess I should try and see what I get when HT is turned off, but then it is turned on, SecurAble (grc.com) reports that I have a 64 bit processor. Deadline should not be an issue, its not till 2008

---

If it\'s an HT processor, and you want to run two models, go to your General prefs on your Account page and set the number of processors to 2. Then you can run 2 models at once.

You\'re right about the deadline not being an issue - there isn\'t any deadline.
The date is there because the server software requires it for other projects.

But he was running Einstein and WCG (though he intends to bow out from them on that machine). The other projects and their unknown cause/effect were the basis of my comment.

That\'s the trouble with shooting in the dark -- with incomplete information, no Message log posted, etc.

---

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

Yes, the messages would be a big help.

If it\'s an HT processor, and you want to run two models, go to your General prefs on your Account page and set the number of processors to 2. Then you can run 2 models at once.

You\'re right about the deadline not being an issue - there isn\'t any deadline.
The date is there because the server software requires it for other projects.

Acutally with 4 CPDN class models on the host and the client only being allowed to use 1 CPU the host is very likely in deadline trouble. Allowing the host to use both CPUs will reduce the pressure but may not reduce it enough to make a noticable difference, 4 models is a large amount of work to do. Don\'t worry too much though, when einsteins deadline gets closer the client will devote enough time to finish it.

---

BOINC WIKI

BOINCing since 2002/12/8