Message boards : BOINC Manager : Boinc always says it's suspended
Message board moderation
Author | Message |
---|---|
Send message Joined: 8 Sep 06 Posts: 5 |
Running this on a Windows XP Pro machine. For some reason it won't go active for more than a few minutes at a time. Here's a part of the log: 9/7/2006 5:52:50 PM||Resuming computation 9/7/2006 5:52:50 PM||Rescheduling CPU: Resuming computation 9/7/2006 5:52:50 PM||Resuming network activity 9/7/2006 5:52:50 PM|SETI@home|Restarting task 12my06aa.5669.66481.667316.3.69_3 using setiathome_enhanced version 515 9/7/2006 5:52:50 PM|SETI@home|Resuming task 12my06aa.5669.66720.47148.3.40_2 using setiathome_enhanced version 515 9/7/2006 5:52:52 PM||Suspending computation - user is active 9/7/2006 5:52:52 PM|SETI@home|Pausing task 12my06aa.5669.66481.667316.3.69_3 (removed from memory) 9/7/2006 5:52:52 PM|SETI@home|Pausing task 12my06aa.5669.66720.47148.3.40_2 (removed from memory) 9/7/2006 5:52:52 PM||Suspending network activity - user is active 9/7/2006 6:02:53 PM||Resuming computation 9/7/2006 6:02:53 PM||Rescheduling CPU: Resuming computation 9/7/2006 6:02:53 PM||Resuming network activity 9/7/2006 6:02:53 PM|SETI@home|Restarting task 12my06aa.5669.66481.667316.3.69_3 using setiathome_enhanced version 515 9/7/2006 6:02:53 PM|SETI@home|Resuming task 12my06aa.5669.66720.47148.3.40_2 using setiathome_enhanced version 515 9/7/2006 6:02:55 PM||Suspending computation - user is active 9/7/2006 6:02:55 PM|SETI@home|Pausing task 12my06aa.5669.66481.667316.3.69_3 (removed from memory) 9/7/2006 6:02:55 PM|SETI@home|Pausing task 12my06aa.5669.66720.47148.3.40_2 (removed from memory) 9/7/2006 6:02:55 PM||Suspending network activity - user is active Any ideas on why this would happen? |
Send message Joined: 29 Aug 05 Posts: 15561 |
9/7/2006 6:02:55 PM||Suspending network activity - user is active What are the settings at your Seti general preferences set to, specifically these two: Do work while computer is in use? Do work only after computer is idle for X minutes. If it's no and a number (default 3 minutes), your computer will wait 3 minutes checking if there are no keys being pressed, mouse being moved or many CPU cycles used, before BOINC allows Seti to crunch on. As soon as you hit the keyboard, move the mouse, BOINC will suspend computations again. Although it's also reported that a videochip on the motherboard can cause this, when you try to run the screen saver. Since the videochip is essentially controlled by the CPU, it can be detected as activity, thus BOINC suspends crunching at that moment. |
Send message Joined: 8 Sep 06 Posts: 5 |
Sorry, should have mentioned that I am running on several machines and this is the only one with problems, so it can't be a general setting. And the log piece I posted was at a time when the machine hadn't been touched by anyone for over an hour. On board video though... that's possible, but BOINC is the screensaver. Shouldn't that fix itself? If not, is there a workaround for this? |
Send message Joined: 29 Aug 05 Posts: 15561 |
If it is the on board video that's causing this, then the only work around is to disable the screen saver in Windows (none), or set it to a Windows screen saver. To say that BOINC is only the screen saver is a bit off though. BOINC is the Managing program, it does quite a bit more than just offer pretty pictures. Besides, the Seti screen saver is built in in the Seti Enhanced science application. Applications that don't have a screen saver will show the standard "bouncing BOINC block". ;-) Oh and there are some bug fixes in the upcoming version 5.6.x of BOINC that fix some screen saver problems. You can wait for that one. |
Send message Joined: 8 Sep 06 Posts: 5 |
So a built into Windows screen saver will NOT cause it to think the CPU is being used? Ok, I'll try changing it to one of them and see what happens. If that doesn't work, I'll try setting it to none. I'll be back tomorrow to let you know what happens. Oh - I didn't mean to imply that BOINC was "just" a screensaver, and I really shouldn't have lumped Seti in with it in the first place, but too late now :) |
Send message Joined: 29 Aug 05 Posts: 15561 |
So a built into Windows screen saver will NOT cause it to think the CPU is being used? If it's one of the non-3D ones, yeah. The Seti screen saver is OpenGL 3D, pretty heavy on the GPU as well. But a GPU on an external board won't take CPU cycles away from the CPU... ;-) |
Send message Joined: 8 Sep 06 Posts: 5 |
I set it to the Windows "mystify" screensaver and it is now working correctly. I don't think I'm willing to add a video card just for the seti screen saver, but I will miss it on that machine. Before I changed to a new motherboard a while back (with onboard video of course) it worked just fine. Oh well. Thanks for the help! |
Send message Joined: 29 Aug 05 Posts: 15561 |
You could set the preference "Do work while computer is in use?" to Yes. That way BOINC won't go idle as soon as the Seti screen saver kicks in. The CPU will give less cycles to Seti, but get it done anyway. (When you do, make sure you save the preferences on the website with the button at the bottom, then open Boinc Manager, Projects tab, select Seti, press Update.) Or you can use no screen saver at all and just power down the monitor. That's what I do, always. :-) If you pass by on the Seti forums, at Number Crunching, there is this post by KWSN - Chicken of Angnor with links to an optimized Seti application. It is optimized to use the SSE parts of your CPU, so the crunching goes quicker. These optimized apps lack the graphics libraries, so you can't see the screen saver even if you wanted to. There are options enough, above putting in a cheap PCI videocard with 32MB of RAM. ;-) |
Send message Joined: 8 Sep 06 Posts: 5 |
I don't want it running while the computer is in use though, on any of the machines I have it running on, so I can't change that setting. I'll look into the optimized version though - as long as it isn't doing graphics anyway, I might as well make it go as fast as it can. As for the video card, well, it's a work computer, so that's just not going to happen. Thanks again for all the help. |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.