| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Linux x64 dual core query |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 11
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have ubuntu 7.04 x64 server with gnome installed running on an amd64 x2 3800 system, I had some problems getting fa@h to work, and with the help of the advisors here got it working by updating boinc from synaptic version 5.4.11 to the getdeb version 5.8.17 and installing ia32libs through synaptic, however, I have noticed since then, instead of running 2 cores at 96-100% each (for a sum total of 192-200%), my system is now running 2 cores at a sum of 100%, ie 10% + 90%, fluctuating so the total is always 100%, never running both at 100%.
I was wondering if this is due to the way the updated boinc client uses resources, or something to do with the ia32libs, and if there is a way to get both cores running at 100% again. the statistics page seems to be about the same, average results/run time ... maybe a little lower. Any help or advice much appreciated. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Very possibly u use the "Minimum Impact" WCG profile. Change it to "Maximum Performance" and all should run all the time using all spare cycles.
----------------------------------------Added: If that, go to My Grid > Device Manager > Device Profile and choose the one attached to your device which is usually 'Default' when starting crunching.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Jul 10, 2007 12:19:55 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I use a custom profile, but it's pretty much maximum performance, since I'm not using the system for anything else right now.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello rogue13,
Another possibility is that your profile only allows BOINC enough RAM to execute one work unit at a time, shifting the other to swap space. Whatever the problem is, at least it sounds like a profile problem. Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Have you set up top 100 percent for memory and resource? If you switch from Mimimum to Custom, it initially assumes the 'Minimum Impact' values. To get the Maximum Performance settings into Custom, first save with 'Maximum', then switch to 'Custom'
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I've changed to maximum, and will see if that makes any difference, my settings were at 100% cpu and 80% memory (I have 768mb RAM available) and system monitor shows 430.5 used out of 751.5 (57.3%) with only slight fluctuations ... I've just checked using a termial and "top", rather than gnomes system monitor, and for most of the time there is only 1 instance of boinc listed running ~99% cpu, with occasional flashes of a second at 1-2%, where there used to be 2 at 99%.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The project should run at lowest priority. Does the System Idle process run in place of one of the work units? If so, something is preventing a work unit from running. The profile is the only thing that should do that.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Update: I just noticed the due date on 1 of the 2 hpf2 tasks running, it's 2 days older than it should be, and at 94% complete, isn't changing, I suspended it and boinc shifted to next task ... 100%/100% again
Looks like the task hung and "locked out" 1 of the processors :S |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Yes, it's know and rare event.... sometimes HPF2 loops ad infinitum. Has been reported to the programmers. A simple suspend of WCG as a project (in Project Tab of the BOINCmgr GUI) and resume will kick it in gear.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ouch!
Yes, we have never tracked down that HPF2 bug. Occasionally it gets stuck. We have found that when that happens, suspending the project then unsuspending it causes it to start from the last checkpoint, after which it goes to completion. Apparently it overwrote something important and actually froze on your system. Normally it just runs without accomplishing anything but still shows active computation, unlike your case, which is why we did not think of that possibility. Glad to hear you found it. Sorry it was due to a persistent bug that we have never been able to track down and squash. Lawrence |
||
|
|
|