Hi,

If you go to the beginning of the BOINC Manager message log and copy paste the first 25 lines to where computing resumes, sample below, we can do a basic read of your configuration. Large jobs like FAAH of recent have high momentary system demands. Just want to make sure it's not your setup.

RICE will run 8 CPU hours, from batch R00200 9 CPU hours, regardless of the device power. That's for many part time crunchers a bonus.

ttyl

29/10/2008 11:46:15||Starting BOINC client version 6.2.25 for windows_intelx86
29/10/2008 11:46:15||log flags: task, file_xfer, sched_ops, checkpoint_debug
29/10/2008 11:46:15||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
29/10/2008 11:46:15||Running as a daemon
29/10/2008 11:46:15||Data directory: D:\Mijn Documenten\BOINCData
29/10/2008 11:46:15||Running under account boinc_master
29/10/2008 11:46:17||Processor: 2 GenuineIntel Intel(R) Core(TM)2 CPU T5600 @ 1.80GHz [x86 Family 6 Model 15 Stepping 6]
29/10/2008 11:46:17||Processor features: fpu tsc pae nx sse sse2 mmx
29/10/2008 11:46:17||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
29/10/2008 11:46:17||Memory: 1.50 GB physical, 2.85 GB virtual
29/10/2008 11:46:17||Disk: 47.98 GB total, 5.58 GB free
29/10/2008 11:46:17||Local time is UTC +1 hours
29/10/2008 11:46:17|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 95711; location: work; project prefs: work
29/10/2008 11:46:17||General prefs: from World Community Grid (last modified 29-Jul-2008 18:31:51)
29/10/2008 11:46:17||Computer location: work
29/10/2008 11:46:17||General prefs: using separate prefs for work
29/10/2008 11:46:17||Reading preferences override file
29/10/2008 11:46:17||Preferences limit memory usage when active to 1380.69MB
29/10/2008 11:46:17||Preferences limit memory usage when idle to 1380.69MB
29/10/2008 11:46:17||Preferences limit disk usage to 5.35GB
29/10/2008 11:47:16|World Community Grid|Restarting task X0000052741460200507050859_0 using hcc1 version 606
29/10/2008 11:47:16|World Community Grid|Restarting task X0000052800872200507051100_1 using hcc1 version 606
29/10/2008 11:52:38||Suspending computation - user request
29/10/2008 11:53:15||Resuming computation

Those parms look healthy enough to tackle any current WCG task.

After 76 million results one would expect to be running a proven piece of software[FAAH]. That said, when there are problems, it's often the graphics part. Disable that and the ss and issues have gone away. You could test that if you want and do check if your display card drivers are the latest.

cheers

[Edit: Clarified the 76 million relating specifically to FAAH]

Hi Tom.
Have you tried to find error messages which would have been produced before Boinc collapses?
If not this is how to proceed.
- Go to your Boinc Data directory.
- Copy the file stdoutdae.txt to a temp/work directory.
- Look at this copy with your favorite editor and search for "client version".
That will find the start of each session still available in that file.
- See if you find unusual error messages in the preceding lines, i.e. at the end of the previous session.

Something else that you could do is to activate the logging of checkpoints. It seems that you have at least two different problems, i.e. FAAH jobs killing Boinc, and FAAH not taking checkpoints, unless you tell us that FAAH jobs are failing too early for having ever been able to take a checkpoint.

To activate the logging of checkpoints-
- edit the file cc_config.xml (also in the Boinc Data directory)
- add the following line
<checkpoint_debug>1</checkpoint_debug>
in the log flags section
- save the file
- from the Advanced view of BoincMgr select Advanced>Read configuration file to activate the logging.

Edit: Corrected the name of the file cc_config.xml (and not cc-config.xml)

Cheers. Jean.

OK Tom, no problem. I have not detailed everything (some people more knowledgeable think you see them as idiots in that case) but I can.

First don't be afraid. When I say your favorite editor I mean something like Notepad or Wordpad, nothing more sophisticated. You will need it only for viewing the file and for finding (Ctrl+F) the starting line of every session. These lines are the only ones which contain "client version", so you see the game.

It is normal that the messages shown from the advanced view start with the current session. This is the reason you will have to look in that stdoutdae.txt file which usually contains several past sessions altogether.

The file stdoutdae.txt is in the Boinc data directory which, according to the message log that you have provided, is accessed via the following path:
C:\Documents and Settings\All Users\Application Data\BOINC

If you do not have a temp or work directory of your own for receiving a copy of the file, either create one or simply copy it to the desktop. You will clean it when you do not need it any longer.

To view this copy of the file-
- right click on its icon
- move your mouse down to "Open with"
- select Notepad or WordPad. That's it.

From my own viewpoint prayers are not mandatory, but they cannot hurt.
Good luck. Jean.

You are right that this is what it should mean, but it is not...
In fact Boinc says that when it received the information that it should stop. Whomever or whichever was the sender. So that can be you directly, or the system because you have asked for rebooting or shutting down. Or possibly the system because automatic updates are activated and after some updates have been done an automatic reboot is triggered. Or... who knows what that I am not thinking of right now.

The interesting thing is that if you have this message at least that means that Boinc has not crashed on an error of its own.

Regarding the benchmark error I have seen it sometimes without any annoying consequence. I think it is issued after a benchmark is started if something else happens in Boinc that would make the benchmark worthless. I presume that it is listed as an error because if Boinc were perfect it should not start a benchmark while it is waiting for the answer to a request for work, for example. I don't know if this is a possible situation in the current version but that could theoretically happen if/when various tasks of Boinc are not well coordinated.

Have you been able to activate checkpoints logging as I suggested?

Cheers. Jean.