After a whopping 35 hours at 88%.

I decided to try this again and have had some of these 614 WUs and none of them has taken 4 hours. They click right along like a real program.

Now if the calculations for the "To Completion" would just recover from the monster induced estimates all would be well in the world. They are slowly recovering, go down about 10% per batch. But they have a way to go. Everything was estimated at over 400 hours there for a while. Down to around 100 now.

At least the WUs are running well in ALL projects well now. I am happy.

OK, so now I have had 2 run over 4 hrs (614) and they ended early due to that. they are "pending validation". I do not understand how incomplete WUs can be validated. Seems that this should be an "error" ending.

If these types of endings are acceptable, does this spawn "children"?

I got a long work unit going on over 18 hours and is reading 56.73% completion with 8 1/2 hrs and climbing.

Result Name App Version Number Status Sent Time Time Due /
Return Time CPU Time (hours) Claimed/ Granted BOINC Credit
CMD2_ 0002-RADIA.clustersOccur-TPM1A.clustersOccur_ 2770_ 2-- - In Progress 6/11/09 17:36:48 6/17/09 08:00:48 0.00 0.0 / 0.0
CMD2_ 0002-RADIA.clustersOccur-TPM1A.clustersOccur_ 2770_ 0-- - No Reply 5/28/09 17:25:59 6/11/09 17:25:59 0.00 0.0 / 0.0
CMD2_ 0002-RADIA.clustersOccur-TPM1A.clustersOccur_ 2770_ 1-- 613 Pending Validation 5/28/09 17:25:55 5/31/09 12:34:15 3.65 55.6 / 0.0

I am the replacement unit sent out to cover the no reply.

The thing is I am curious as to what is happening as the pending work unit shows complete with 3.65hrs.

I just turned checkpointing on and the work unit is checkpointing

6/12/2009 6:11:56 PM|World Community Grid|[checkpoint_debug] result CMD2_0002-RADIA.clustersOccur-TPM1A.clustersOccur_2770_2 checkpointed

Also I am not running this unit on a slow machine, it is vertually brand new

6/11/2009 7:08:02 PM||Starting BOINC client version 6.2.28 for windows_intelx86
6/11/2009 7:08:02 PM||log flags: task, file_xfer, sched_ops
6/11/2009 7:08:02 PM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
6/11/2009 7:08:02 PM||Running as a daemon
6/11/2009 7:08:02 PM||Data directory: C:\ProgramData\BOINC
6/11/2009 7:08:02 PM||Running under account boinc_master
6/11/2009 7:08:02 PM||Processor: 8 GenuineIntel Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz [Intel64 Family 6 Model 26 Stepping 4]
6/11/2009 7:08:02 PM||Processor features: fpu tsc pae nx sse sse2 pni mmx
6/11/2009 7:08:02 PM||OS: Microsoft Windows Vista: Home Premium x64 Editon, Service Pack 2, (06.00.6002.00)
6/11/2009 7:08:02 PM||Memory: 11.99 GB physical, 23.91 GB virtual
6/11/2009 7:08:02 PM||Disk: 916.44 GB total, 670.60 GB free
6/11/2009 7:08:02 PM||Local time is UTC -4 hours
6/11/2009 7:08:02 PM|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 913226; location: home; project prefs: default
6/11/2009 7:08:02 PM||General prefs: from World Community Grid (last modified 18-Apr-2009 09:32:22)
6/11/2009 7:08:02 PM||Computer location: home
6/11/2009 7:08:02 PM||General prefs: no separate prefs for home; using your defaults
6/11/2009 7:08:02 PM||Preferences limit memory usage when active to 9208.55MB
6/11/2009 7:08:02 PM||Preferences limit memory usage when idle to 11050.26MB
6/11/2009 7:08:02 PM||Preferences limit disk usage to 18.63GB

Any ideas should I stop this work unit or just let it go and see what happens?

Just as a side note could it be the pending work unit went up to a certain point and time and stopped and that there are now child work units waiting to process and could perhaps my computer intends to complete the entire work unit?
If so how would credit and points be allocated can you even validate a partial work unit against a fully completed work unit?

The 4 CPU hours primary cut off and the secondary 8 hours "pass the bucket" (not heard or seen anyone reporting these), is not set in stone. As the techs learn and with the monsters out of the way by the fastest majority under 6.14, it's not unlikely the hours will be increased to also reduce the server side scheduler load, which is the one doing the toughest part [hands off from the update button please ;-]. Presently WCG aims for project means of 7 hours (see chart 5 here ). HCMD2 sits on 2.62 hours average as of yesterday, slowly creeping up as all the 6.13 units come out of the system. At one time the project ran a mean of under 1 hour.