| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Human Proteome Folding - Phase 2
Thread: Invalid result after 100% completion |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 37
|
|
| Author |
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges: 
|
My latest failure was a 100% break event, it stopped at 100%, but it didn't finish. I'm running WinXP btw.
If you'd like me to test the 6.6 version, I'm willing to. Just let me know what I should do to get it and how to install and such. Bart |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi dividedbymyself,
Sekerob has been fuming about the work units lost by his 6.6 installation. I think it is too bleeding edge to recommend just yet.  Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Was I that loud in the padded back room ;?
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
dividedbymyself
Cruncher Joined: Aug 10, 2008 Post Count: 43 Status: Offline Project Badges:
|
I didn't hear a thing. But I'm wearing my earplugs...
 Anyway, when you need someone to test it on 100% stopping failures, just ask. I won't complain about lost time  Bart |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Another way to stop the client changing project early is to set the project switching time to 0 (zero). So far, not had a single job stop early. Just finishing before changing over to another project. Think those on the very long CPDN might like that having one core just run on and on with their model, possibly only until the client is restarted. That would need testing.
----------------------------------------Note that the local pref seem not to allow zero and default back to 60 minutes, but the WCG website profile certainly allows zero by going to the Custom profile > Cross Project settings > Switch between applications every: 0.00 minutes Clear the local prefs of course. Correction: My beta client shows 60 minutes in the prefs but the override file shows <cpu_scheduling_period_minutes>0.000000</cpu_scheduling_period_minutes> . So much for the Berkeley "You get what you see consistency". I'll report it at their dev forum.  To add: Strange the 6 decimal, for the global prefs only show in whole minutes.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 2 times, last edit by Sekerob at Feb 28, 2009 10:37:52 AM] |
||
|
|
vepaul
Senior Cruncher Belgium Joined: Nov 17, 2004 Post Count: 261 Status: Offline Project Badges: 
|
In fact, I have about the same problem. Although my PC is a recent HP Pavilion running on an Intel Core2 Quad.
The faah run to good results, the dddt usually too, the protein folding 2 take a very long time between completion and a result which practically never is valid. Sometimes 10 of them on the same day end in error. The CEP got better lately, until it stopped altogether. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello vep,
It sounds as though you are running Vista. A number of people have problems with HPF2 on Vista quad-cores. Sekerob cured this on one of his systems by downloading BOINC 6.6, which is very experimental and buggy. For the moment, I just recommend a local preference file that omits HPF2 on any computer that has this problem. It sounds as though it will eventually be fixed and i only seems to affect Vista quad-cores. Lawrence |
||
|
|
|