The science app has *always* been 2 parts, a stager or controller and a worker. One uses no CPU time, few seconds during the job, the other does it all. For CEP2 is a same structure with 3 or 4 parts, depending on platform.

--//--

OS: Win 7 x64
RAM: 4 GB DDR3
CPU: Intel(R) Core(TM) i7-2720QM CPU @ 2.20GHz - HT on
Projects selected: CEP2 max 3 WU, DSFL
Client: 6.12.34 for windows_x86_64


This device only picked up DFSL this morning batch 50 and started running slowly, very un-responsive.

These WUs restart with typical memory size, then start to go to ~1 GB, device physical memory goes to 100%, then the WUs restart over again.

I'll think I'll suspend about half WU until they get processed, I've deselected this project in the mean time.


Here's the last part of my log:


10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0260_1 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0503_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0792_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0977_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_1070_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0274_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_1088_0 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Task DSFL_00000050_0000024_0819_1 exited with zero status but no 'finished' file
10/31/2011 9:05:26 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:05:26 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0260_1 using dsfl version 619
10/31/2011 9:05:26 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0503_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0792_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0977_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_1070_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0274_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_1088_0 using dsfl version 619
10/31/2011 9:05:38 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0819_1 using dsfl version 619
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0260_1 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0503_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0792_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0977_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_1070_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0274_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_1088_0 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Task DSFL_00000050_0000024_0819_1 exited with zero status but no 'finished' file
10/31/2011 9:08:12 AM | World Community Grid | If this happens repeatedly you may need to reset the project.
10/31/2011 9:08:12 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0260_1 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0503_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0792_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0977_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_1070_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0274_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_1088_0 using dsfl version 619
10/31/2011 9:08:25 AM | World Community Grid | Restarting task DSFL_00000050_0000024_0819_1 using dsfl version 619


EDIT: Added client version and spelling.

This issue does not seem to be isolated to batch 50 as I have a couple of machines running batch 49 and they are showing the same high memory usage problem.

From pesonal experience it seems that you need sufficient memory (about 1GB) for each of the DFSL tasks that you have running, and is also dependant on your memory usage settings. I'm currently using the default values of 50% (in use) and 90% (idle) on all my systems, and as long as the system is not in use then all of my DFSL tasks will run. This drops to 50% when the system is in use due to insufficient system memory allocation.

Hopefully this will help a bit, but it doesn't explain why the memory usage has shot through the roof since batch 49.

OS - vista
RAM - 2Gb
core 2 duo
boinc - 6.10.58

The working set size was 232Mb for batch 49 but 874Mb for batch 50.
The batch 50 is a repair unit and started immediately as high priority. BOINC then tried 3 other tasks in the queue but has left them as 'waiting to run, suspended' so is running only one work unit with the second core left idle.

Just notice that there are a series of files in the slots directory with names like "dsfl.target_00000050__1581432322_log.txt" all of which contain messages which start with:

WARNING: The search space volume > 27000 Angstrom^3 (See FAQ)

Could this explain the issues? (I haven't found the referred to FAQ entry yet, so if anyone can help...)