| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 33
|
|
| Author |
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
BOINC does not abort, because the Agent does not know if the DC project will still accept the result. It merely told the agent, it wanted a result back before a certain date.
----------------------------------------In The Case of WCG after exact 7x24 hours an extra copy will be send out to another machine, depending on the urgency it gets a 7 to 1 day deadline. The overdue Work Unit is marked 'No Reply'. If the 'No Reply' copy is returned before the 4th copy, it is accepted with credit. If after 4th copy returned, it gets marked 'Too Late' and no points credited. Yes, the FAAH actual CPU times are spread quite widely. On my C2D seen from 8.25 to 5.00 hours in the last few days (well i did not see it until looking in BOINCview which logs and tabulates all completed work and allows sorting on project and completion times). cheers
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Jan 17, 2007 11:49:33 AM] |
||
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18667 Status: Offline Project Badges:
|
THe part I like best about whatever was changed is the fact that every WU of the same project doesn't have the same TC. While they tend to be reasonably close to each other, they do differ. For me, that implies that potentially (and hopefully) if a new batch of WUs gets loaded that just happens to have noticably different TC, those will show up and not skew the rest of the WUs you may have in your queue.
---------------------------------------- |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Hi, As a result of receiving WUs with a lower TC, i now have > 60 WUs in queue where my connect-settings are on 2.0 (days). In my case with 24 x 7 running machines i will probably make all those WUs before the deadline, but i am sure that this does not count for everyone that experiences the above. Gerard. I just took a look at your machines. The 60 WU's queued is spread across several of your machines. I took a look at each machine and given the average runtimes for the different projects on those machines, each machine had about 48 hours of work queued up. I was actually surprised that it added up so well. Out of curiosity - could you set one of those computers to not receive new work and let us know when it completes all of its work? As Sekerob pointed out, during the transition from mediocre estimates to the better estimates there are going to be some problems since the boinc client itself tries to adapt to the estimates a project provides and thus it has to adapt from medicore estiamtes to better estimates. However, just so everyone keeps this in mind. Our projects are non-deterministic. This means that we cannot predict how long they will take and there can be variation of us much as 50%. |
||
|
|
|