Not running on all threads.

[Rob_B]

I have a problem with my new cruncher. When I got home and checked it, it was only crunching 7 WU's. I had seen it yesterday too, but thought it was just a bug. It crunched 8 WU's again after a restarting Boinc.

But now I didn't. After restarting Boinc it started at 8 and stopped 2 WU's after less than a minute. It shows these as "Waiting to run". But it doesn't show any reason in the event log. After restarting Win7 it is working like it should.



Could this be a temperature problem? It has been hot the last 2 days. If it is, can I see that somewhere? If Win7 is throttling the CPU I would like to see what and why it is doing it. CPU-z showed the CPU was still running at 3.7GHz.
Core temp showed these temps and CPU load.



This is what I could see in the event log. It restarted the SN2S WU twice, but it doesn't say it got suspended before.

simracingmedia2

21			22-5-2012 17:26:41	Reading preferences override file	
22			22-5-2012 17:26:41	Preferences:	
23			22-5-2012 17:26:41	   max memory usage when active: 1895.81MB	
24			22-5-2012 17:26:41	   max memory usage when idle: 3412.47MB	
25			22-5-2012 17:26:41	   max disk usage: 29.25GB	
26			22-5-2012 17:26:41	   (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)	
27			22-5-2012 17:26:41	Not using a proxy	
28	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2008_02008904_0 using cfsw version 605 in slot 5	
29	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2008_02008181_0 using cfsw version 605 in slot 7	
30	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2009_02009271_1 using cfsw version 605 in slot 6	
31	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2012_02012715_0 using cfsw version 605 in slot 2	
32	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2014_02014457_0 using cfsw version 605 in slot 3	
33	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_1888_01888983_1 using cfsw version 605 in slot 0	
34	World Community Grid	22-5-2012 17:26:42	Restarting task cfsw_2015_02015364_0 using cfsw version 605 in slot 4	
35	World Community Grid	22-5-2012 17:26:46	Computation for task cfsw_1888_01888983_1 finished	
36	World Community Grid	22-5-2012 17:26:46	Restarting task cfsw_2016_02016538_0 using cfsw version 605 in slot 1	
37	World Community Grid	22-5-2012 17:26:48	Started upload of cfsw_1888_01888983_1_0	
38	World Community Grid	22-5-2012 17:26:54	Finished upload of cfsw_1888_01888983_1_0	
39	World Community Grid	22-5-2012 17:27:15	task cfsw_1888_01888983_1 aborted by user	
40	World Community Grid	22-5-2012 17:27:16	Restarting task SN2S_q01137_0000031_0179_2 using sn2s version 613 in slot 8	
41	World Community Grid	22-5-2012 17:27:58	update requested by user	
42	World Community Grid	22-5-2012 17:27:59	Sending scheduler request: Requested by user.	
43	World Community Grid	22-5-2012 17:27:59	Reporting 1 completed tasks, requesting new tasks for CPU	
44	World Community Grid	22-5-2012 17:28:01	Scheduler request completed: got 2 new tasks	
45	World Community Grid	22-5-2012 17:28:03	Started download of cfsw_2083_02083513_A02083513.sql	
46	World Community Grid	22-5-2012 17:28:03	Started download of cfsw_2083_02083642_A02083642.sql	
47	World Community Grid	22-5-2012 17:28:04	Finished download of cfsw_2083_02083513_A02083513.sql	
48	World Community Grid	22-5-2012 17:28:04	Finished download of cfsw_2083_02083642_A02083642.sql	
49	World Community Grid	22-5-2012 17:29:06	Restarting task SN2S_q01137_0000031_0179_2 using sn2s version 613 in slot 8	
50	World Community Grid	22-5-2012 17:31:10	Computation for task cfsw_2008_02008904_0 finished	
51	World Community Grid	22-5-2012 17:31:10	Starting task cfsw_2017_02017113_0 using cfsw version 605 in slot 0	
52	World Community Grid	22-5-2012 17:31:12	Started upload of cfsw_2008_02008904_0_0

I have to get a better cooler for this 3770, no matter if this is a temp problem or not.
Does any of you know a good low profile cooler that can cool this CPU and is quiet? Max 82 mm high.

Thanks

[hendermd]

Is your vcore set to auto? For both my Sandy and Ivy when trying to find the lowest vcore while running prime95, 1 thread would drop when voltage was to low.

[NKrader]

are you sure its not running other tasks? my arima will do this have like 30+ projects started, but only running 16 at a time.. check the list and make sure there arent others going, sense you couldnt see this by that you have it sorted by progress. sort by status.

ive done that before. if its not that, hmmmmm.

[Plan.B]

maroon user name nomination for hendermd

Is your vcore set to auto? For both my Sandy and Ivy when trying to find the lowest vcore while running prime95, 1 thread would drop when voltage was to low.

[Rob_B]

Is your vcore set to auto? For both my Sandy and Ivy when trying to find the lowest vcore while running prime95, 1 thread would drop when voltage was to low.

Vcore is at auto. The only other options I have are, +0.05v, +0.10v and +0.15v

are you sure its not running other tasks? my arima will do this have like 30+ projects started, but only running 16 at a time.. check the list and make sure there arent others going, sense you couldnt see this by that you have it sorted by progress. sort by status.

This is on a fresh Win7 install and with only drivers installed for the chipset and LAN.
When I look at the task manager it show 7 WCG WU's at 13% each and all other tasks are at 0%

[D_A]

BOINC Client will sometimes stop units if it's run out of system RAM based on settings. Check how much you have allowed to use and adjust it up a few % points if needed.

[Johnmark]

I've seen this a couple of times myself.
No your temps. are not too hot unless you are getting errors. Throttling should not happen until your up near 100c.
Your absolutely sure another unit isn't running somewhere else on the pages of pending work to be done? Yea I didn't see that one that was running pages down on the screen.
Have you had a lot of WU's turn up bad? my machine once returned a whack of bad WU's and boinc would not run all the WU's it was supposed to for a while, less than 24hrs is all I recall.

[Rob_B]

BOINC Client will sometimes stop units if it's run out of system RAM based on settings. Check how much you have allowed to use and adjust it up a few % points if needed.

I had it at default settings. 50% when computer is used and 90% when idle. Changed it to 90% when in use.
2.7 GB of the 4 GB is used. Does CFSW use that much more memory than the other projects?

I've seen this a couple of times myself.
No your temps. are not too hot unless you are getting errors. Throttling should not happen until your up near 100c.
Your absolutely sure another unit isn't running somewhere else on the pages of pending work to be done? Yea I didn't see that one that was running pages down on the screen.
Have you had a lot of WU's turn up bad? my machine once returned a whack of bad WU's and boinc would not run all the WU's it was supposed to for a while, less than 24hrs is all I recall.

I'm sure only 7 WU's are running. I use BoincTask to monitor the crunchers and that shows every thing on the screen without having to scroll down.
Only one error and that happened when I restarted Boinc. But I see that I have 23 inconclusive WU's for this PC.

I'm going to see if the memory setting helped. If it doesn't I will try an other project.

[Rob_B]

Thanks all. Looks like D_A got the right answer for this one.

After changing the memory settings in Boinc it has been crunching on 8 threads all the time. Thanks D_A.

Now lets see how long it takes to make it on the top hosts list.

[my87csx481]

Just to share something I've witnessed....

I have a 2600K with 4GB of RAM. Every once in a while, I peek at it and it's up near 3.8GB used ram, and seems less than responsive. If I reboot, it's at ~2.7GB used ram. This is crunching CFSW.

Thanks for this thread. I had forgotten all about needing to change the projects on that machine, or get around to putting in a larger ram kit that's just sitting around. Maybe just bumping the percentages will work.

Glad you got your issue sorted out.