| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home
Thread: this is a really long work unit |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 149
|
|
| Author |
|
|
pacerintl
Cruncher USA Joined: Nov 7, 2006 Post Count: 47 Status: Offline Project Badges: 
|
I've had quite a few of these jobs run long too...one ran over ninety hours and just before finishing there is a computational error...and I get 0 credit for wasting over 90 hours. This has happened a few times, sometimes I get partial credit...sometimes nothing. Should I just kill jobs that are running over 20hours ?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello pacerintl,
knreed runs a script every few days to award credit to long FAAH units that have timed out. Hasn't he caught any of your over-long units? Lawrence |
||
|
|
mikaado
Cruncher Joined: Dec 3, 2007 Post Count: 14 Status: Offline |
Should I just kill jobs that are running over 20hours ? I had two of these monster WU's and the another one went all right. But the other was 98 % in 42h. Then I shutted down my computer and now it started to calculate it from the start. Well, let's see how far it goes this time  |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
hmmm, that's no good,
----------------------------------------The recommended shut-down if working on monster tasks in particular is, to open BOINCmgr Advanced View and go into the Advanced menu. Choose the Shut Down Connected Client.... The reason for this option rather than exiting BOINC directly is, that it will stop the core client (BOINC.exe) too even when running as service, provided you have permission. Why manually shutting down?: Because some OSses could be too quick and not allow BOINC to write the checkpoint save file completely from where a task could resume. Particular Vista with pre 5.10.35 BOINC versions suffer this problem, to include the risk of job corruption and complete loss. ciao Note to self: A candidate for a short and separate FAQ Index item and cross reference to the Checkpoint Save article.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
mikaado
Cruncher Joined: Dec 3, 2007 Post Count: 14 Status: Offline |
This is the first time I have any problems with BOINC client. I'm running 5.10.45 client with vista x64. But I think that the shutdown might be the issue. Just thought tech's might like to know.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
This is the first time I have any problems with BOINC client. I'm running 5.10.45 client with vista x64. But I think that the shutdown might be the issue. Just thought tech's might like to know. That's what I thought, and know.... follow the link in the Checkpoint FAQ which has at bottom a link to the Vista FAQ. Data location is important and should NOT be in the C:\Program Files structure. BOINC 6 will put it in the C:\ProgramData structure Very large files require substantial time to be written and closed, so experience showed even a forced delay of 60 seconds was not enough for some.... hence the manual shutdown recommendation.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So what should I do with queued work that will no way make the deadline?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Questar,
knreed extended the deadline some time ago. If you already had these units, then the server knows about the new deadline, even if your client does not. If these are recent work units that you cannot complete, then go ahead and abort them. Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the reply.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Credit has now been awarded to all results for FightAIDS@Home that were running the 'monster' workunits. If the cpu_time reported was more than 10,000 seconds, then credit and runtime were awarded based upon what the computer claimed and reported. If the cpu time reported was less than 10,000 seconds, then credit and runtime were awarded based upon the averages of 550 BOINC Credits and a runtime of 230699 seconds (~64.1 hours). This second case was just run for the first time. We awarded the following: BOINC Credit: 239250 WCG Points: 1674750 Results: 435 Run Time: 100354065 seconds (3.2 years) Hosts: 348 Users: 284 Can we assume you will be running this again? I've got work units aborting all over the place and reporting 0 time. Thanks. |
||
|
|
|