Posts by jrlecker
log in
1) Message boards : Number crunching : Excessive Work Units Running (Message 58875)
Posted 15 Jun 2013 by jrlecker
As for your problem I still think it is your settings:

If it was preferences, than why don't any of my 13 other projects have issues?


On the first line you are telling Boinc to only use 4 cpu's max, then on the next set you are telling it to use upto 80% of the 8 cores available on your i7.


Uh, no. I don't have 8. This machine only has 6. But it really doesn't matter because the first option isn't used on new versions. I needed both options set for multiple computers with all different versions installed.


I think this needs some more explanation for me to understand better, especially on how you can have 7 wu's running at the same time if you have the max processors set to 4? But as for the 'normal' or 'low' priority settings, this is set by Boinc and could be a mislabeling in your thinking. 'Normal' may mean 'normal to Boinc', but mean 'low priority' to the machine. My task manager does not show normal or any other kind of priority, where do you see that?


Wrong again. I opened up the system manager to see why the computer wasn't responding. It's computer normal priority. That's why none of my other normal priority computer programs would respond.
Also, look my event log. 7 new processes start with no processes ending. There was 11 total units running. The 4 that was suppose to plus the 7 from here.


Unfortunately this happens in Boinc sometimes, suspending a unit does NOT always truly suspend it. I don't know if waiting a couple of minutes fixes it or if you just have to physically stop Boinc altogether, I have never waited long enough to find out. It COULD be doing a checkpoint, I just don't know, but ALOT of people have the same problem. Some projects have checkpoints in the 20mb size range, so they can take a while to write.


Yes, it may not be automatic, but out of the 14 projects I'm running, they all respond within seconds except this. The one task unit went from 8% to 45% after some 3-4 minutes and the others were just as bad. Sorry, that tells me it's broken.
2) Message boards : Number crunching : Excessive Work Units Running (Message 58838)
Posted 13 Jun 2013 by jrlecker
I am burning up GPUs because of the location of the machine. The environment prohibits me from running things at 100% and since there is no percentage restriction preference it's either an all or nothing deal. All burns out hardware so I have to go with nothing.

Having it write every 5 seconds is fine. I normally have so much other stuff writing to my hard drives constantly that this isn't an issue. Though I could probably boost that up now that I don't have some of the other machines that needed it.

50% ram for projects is never a factor. In fact I could easily cut it in half with no negative affect. BOINC almost never goes above 10% unless I'm running 3 or more climate units at once. Yesterday was just basic forum writing when I experienced the problem because I had just logged back into the system and hadn't started a lot of stuff, but most of the time there will be a lot more intensive usage. This is also why I need the CPU restriction in place.

There's also a network restriction in place where it only connects once a day for like 20 minutes or so to not interfere so much with other things, and at least one or two times a week it will miss that connection period, so the network settings have to stay as they are too. I'm probably going to end up upping from 0.5 to 1.5 and 1.5 to 2 or higher. It will create some havoc for a few days but level out so that I actually have enough units all the time.

Selecting or un-selecting the site's send GPU units is irrelevant. When the BOINC manager has no demand for those units (as it's disabled) it will never ask sites for them. Every once in a while I can enable it for a short time so it needs to be able to ask for units.



Now all this is fine and dandy but it hasn't addressed the issue regarding why this program's units don't adhere to configuration preferences in the first place.
3) Message boards : Number crunching : Excessive Work Units Running (Message 58808)
Posted 13 Jun 2013 by jrlecker
No errors in the event log (shown below). It shows that immediately after the units were downloading they started and no currently running units were stopped. At the moment they kicked in, I was typing up some documentation on another message board (so minimal resources were in use on other computer processes). My computer's task manager had 7 instances of milky..(rest of filename that I don't remember).exe process running at normal priority (which reminds me, it should be set to a low priority) and their combined CPU usage equaled 100% across my entire system. All of my other running processes were 0 including the other BOINC project units that were set to running status. When I suspended the work unit, the CPU process never changed, and the work unit % completed continued to go up. Also suspending the entire project had no effect either. I had to suspend BOINC altogether before I think it finally stopped.

Another funny thing I just noticed... I have a work unit right now that says it needs 4 CPUs to process. I have an open slot for running units (only 3 work units are running). (Weird the manager hasn't gone out to other projects to request work, but I think it sees I have units that aren't running so it assumes I have enough work) Anyways, that's probably never going to happen because of all of the other units with high priority status. But what's funny is that if I increase my computer output to 5 CPUs available, the unit changes to saying wait for 4 CPUs to wait for 5 CPUs. Uh... huh??



My manager preferences: (I copied out of my prefs page on the website since I could copy/paste that, but I double checked them against the actual manager config and they are the same)

Suspend work when non-BOINC CPU usage is above
0 means no restriction
Enforced by version 6.10.30+ 25%
Switch between tasks every
Recommended: 60 minutes 120 minutes
On multiprocessors, use at most 4 processors
On multiprocessors, use at most
Enforced by version 6.1+ 80% of the processors
Use at most
Can be used to reduce CPU heat 75% of CPU time
Disk and memory usage
Disk: use at most 50 GB
Disk: leave free at least
Values smaller than 0.001 are ignored 1.5 GB
Disk: use at most 50% of total
Tasks checkpoint to disk at most every 5 seconds
Swap space: use at most 95% of total
Memory: when computer is in use, use at most 50% of total
Memory: when computer is not in use, use at most 90% of total

Also, not sure if it's relevant or not, but my GPU is always disabled. The moment I enable it, the thing runs at 100%, overheats and burns out. I've had to replace it twice as it's literally burned up the hardware because of this program so I've just disabled that forever



6/11/2013 6:33:28 AM | | Resuming after OS suspension
6/11/2013 3:52:59 PM | | Resuming computation
6/11/2013 3:52:59 PM | | Resuming network activity
6/11/2013 3:53:02 PM | | Windows is resuming operations
6/11/2013 3:53:02 PM | FreeHAL@home | Fetching scheduler list
6/11/2013 3:53:18 PM | LHC@home 1.0 | Sending scheduler request: Requested by project.
6/11/2013 3:53:18 PM | LHC@home 1.0 | Not reporting or requesting tasks
6/11/2013 3:53:20 PM | LHC@home 1.0 | Scheduler request completed
6/11/2013 3:53:30 PM | | Project communication failed: attempting access to reference site
6/11/2013 3:53:32 PM | | Internet access OK - project servers may be temporarily down.
6/11/2013 4:19:30 PM | NumberFields@home | Sending scheduler request: To fetch work.
6/11/2013 4:19:30 PM | NumberFields@home | Requesting new tasks for CPU
6/11/2013 4:19:33 PM | NumberFields@home | Scheduler request failed: Couldn't connect to server
6/11/2013 4:19:39 PM | Milkyway@Home | Sending scheduler request: To fetch work.
6/11/2013 4:19:39 PM | Milkyway@Home | Requesting new tasks for CPU
6/11/2013 4:19:41 PM | Milkyway@Home | Scheduler request completed: got 12 new tasks
6/11/2013 4:19:43 PM | Milkyway@Home | Started download of milkyway_nbody_1.18_windows_x86_64__mt.exe
6/11/2013 4:19:43 PM | Milkyway@Home | Started download of libgomp_64-1_nbody_1.18.dll
6/11/2013 4:19:44 PM | Milkyway@Home | Finished download of libgomp_64-1_nbody_1.18.dll
6/11/2013 4:19:44 PM | Milkyway@Home | Started download of pthreadGC2_64_nbody_1.18.dll
6/11/2013 4:19:45 PM | Milkyway@Home | Finished download of milkyway_nbody_1.18_windows_x86_64__mt.exe
6/11/2013 4:19:45 PM | Milkyway@Home | Finished download of pthreadGC2_64_nbody_1.18.dll
6/11/2013 4:19:45 PM | Milkyway@Home | Started download of milkyway_separation_1.02_windows_x86_64__opencl_nvidia.exe
6/11/2013 4:19:45 PM | Milkyway@Home | Started download of nbodylua_EMD_1.18_10K.lua
6/11/2013 4:19:47 PM | Milkyway@Home | Finished download of nbodylua_EMD_1.18_10K.lua
6/11/2013 4:19:47 PM | Milkyway@Home | Started download of nodark_10K_fixed.hist
6/11/2013 4:19:47 PM | | Project communication failed: attempting access to reference site
6/11/2013 4:19:48 PM | | Internet access OK - project servers may be temporarily down.
6/11/2013 4:19:48 PM | Milkyway@Home | Finished download of milkyway_separation_1.02_windows_x86_64__opencl_nvidia.exe
6/11/2013 4:19:48 PM | Milkyway@Home | Finished download of nodark_10K_fixed.hist
6/11/2013 4:19:48 PM | Milkyway@Home | Started download of Dark_Test.hist
6/11/2013 4:19:48 PM | Milkyway@Home | Started download of 79_constrained_rev_3.prmtrs
6/11/2013 4:19:48 PM | Milkyway@Home | Starting task de_separation_20_2s_sscon_1_1370980288_1181_0 using milkyway version 102 (opencl_nvidia) in slot 4
6/11/2013 4:19:48 PM | Milkyway@Home | Starting task de_separation_20_2s_sscon_1_1370980288_1179_0 using milkyway version 102 (opencl_nvidia) in slot 5
6/11/2013 4:19:48 PM | Milkyway@Home | Starting task de_separation_20_2s_sscon_1_1370980288_1180_0 using milkyway version 102 (opencl_nvidia) in slot 6
6/11/2013 4:19:49 PM | Milkyway@Home | Finished download of Dark_Test.hist
6/11/2013 4:19:49 PM | Milkyway@Home | Finished download of 79_constrained_rev_3.prmtrs
6/11/2013 4:19:49 PM | Milkyway@Home | Started download of 79_DR_8_rev_1.stars
6/11/2013 4:19:49 PM | Milkyway@Home | Started download of nbodylua_CHISQ_1.12_10K.lua
6/11/2013 4:19:50 PM | Milkyway@Home | Finished download of nbodylua_CHISQ_1.12_10K.lua
6/11/2013 4:19:50 PM | Milkyway@Home | Started download of p-21-2s-sscon.txt
6/11/2013 4:19:52 PM | Milkyway@Home | Finished download of p-21-2s-sscon.txt
6/11/2013 4:19:52 PM | Milkyway@Home | Started download of stars-21-sansSgr.txt
6/11/2013 4:19:54 PM | Milkyway@Home | Finished download of 79_DR_8_rev_1.stars
6/11/2013 4:19:54 PM | Milkyway@Home | Finished download of stars-21-sansSgr.txt
6/11/2013 4:19:54 PM | Milkyway@Home | Starting task ps_separation_79_DR8_rev_3_1370980288_1174_0 using milkyway version 102 (opencl_nvidia) in slot 7
6/11/2013 4:19:54 PM | Milkyway@Home | Starting task de_separation_79_DR8_rev_3_1370901800_441110_2 using milkyway version 102 (opencl_nvidia) in slot 8
6/11/2013 4:19:54 PM | Milkyway@Home | Starting task ps_separation_79_DR8_rev_3_1370980288_1175_0 using milkyway version 102 (opencl_nvidia) in slot 9
6/11/2013 4:19:54 PM | Milkyway@Home | Starting task de_separation_21_2s_sscon_1_1370980288_1182_0 using milkyway version 102 (opencl_nvidia) in slot 10
6/11/2013 4:21:40 PM | | Suspending network activity - time of day
6/11/2013 4:21:47 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1179_0 suspended by user
6/11/2013 4:21:51 PM | Milkyway@Home | task de_separation_21_2s_sscon_1_1370980288_1182_0 suspended by user
6/11/2013 4:21:55 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1181_0 suspended by user
6/11/2013 4:21:59 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1175_0 suspended by user
6/11/2013 4:22:02 PM | Milkyway@Home | task de_separation_79_DR8_rev_3_1370901800_441110_2 suspended by user
6/11/2013 4:22:22 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1174_0 suspended by user
6/11/2013 4:22:32 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1180_0 suspended by user
6/11/2013 4:23:16 PM | Milkyway@Home | project suspended by user
6/11/2013 4:23:45 PM | | Suspending computation - user request
6/11/2013 4:24:14 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1180_0 aborted by user
6/11/2013 4:24:15 PM | Milkyway@Home | Computation for task de_separation_20_2s_sscon_1_1370980288_1180_0 finished
6/11/2013 4:24:19 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1179_0 aborted by user
6/11/2013 4:24:21 PM | Milkyway@Home | Computation for task de_separation_20_2s_sscon_1_1370980288_1179_0 finished
6/11/2013 4:24:26 PM | Milkyway@Home | task de_separation_21_2s_sscon_1_1370980288_1182_0 aborted by user
6/11/2013 4:24:27 PM | Milkyway@Home | Computation for task de_separation_21_2s_sscon_1_1370980288_1182_0 finished
6/11/2013 4:24:34 PM | Milkyway@Home | task de_separation_20_2s_sscon_1_1370980288_1181_0 aborted by user
6/11/2013 4:24:35 PM | Milkyway@Home | Computation for task de_separation_20_2s_sscon_1_1370980288_1181_0 finished
6/11/2013 4:24:39 PM | | Resuming computation
6/11/2013 4:24:45 PM | Milkyway@Home | task de_separation_79_DR8_rev_3_1370901800_441110_2 aborted by user
6/11/2013 4:24:46 PM | Milkyway@Home | Computation for task de_separation_79_DR8_rev_3_1370901800_441110_2 finished
6/11/2013 4:24:47 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1174_0 resumed by user
6/11/2013 4:25:00 PM | Milkyway@Home | project resumed by user
6/11/2013 4:25:01 PM | Milkyway@Home | Resuming task ps_separation_79_DR8_rev_3_1370980288_1174_0 using milkyway version 102 (opencl_nvidia) in slot 7
6/11/2013 4:32:52 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1174_0 aborted by user
6/11/2013 4:32:53 PM | Milkyway@Home | Computation for task ps_separation_79_DR8_rev_3_1370980288_1174_0 finished
6/11/2013 4:35:52 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1175_0 resumed by user
6/11/2013 4:35:52 PM | Milkyway@Home | Resuming task ps_separation_79_DR8_rev_3_1370980288_1175_0 using milkyway version 102 (opencl_nvidia) in slot 9
6/11/2013 4:35:54 PM | Milkyway@Home | task ps_separation_79_DR8_rev_3_1370980288_1175_0 aborted by user
6/11/2013 4:35:55 PM | Milkyway@Home | Computation for task ps_separation_79_DR8_rev_3_1370980288_1175_0 finished
6/11/2013 5:01:18 PM | NFS@Home | Computation for task G3p706_687378_1 finished
4) Message boards : Number crunching : Excessive Work Units Running (Message 58729)
Posted 11 Jun 2013 by jrlecker
Why is it setup that work units can override any settings and run outside of my program preferences. I have it setup to only use 4 CPUs at a max of 75% of total processing power. All of a sudden today, my computer went to lag city because over half a dozen units just started working on top of the 4 other work units from other projects. Took me 5 minutes to cancel them just so my computer would respond at a somewhat decent rate again.

Suspending the unit did absolutely nothing. They just continued to burn up CPU power in the background. It's like the program ignored anything I tried to do to it to stop it. I basically had to abort the process to kill them.

I don't mind running units, but I have 14 projects I am running on this machine and I can't have this project running 100% of all my processors. There is a queue, and these work units need to wait their turn like all the others.

I guess that would explain some my lower output levels on other projects too.

Need to get this fixed NOW!




Main page · Your account · Message boards


Copyright © 2017 AstroInformatics Group