| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Discovering Dengue Drugs - Together - Phase 2 Forum
Thread: Grossly overestimated run time |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
cw64
Advanced Cruncher Joined: Oct 6, 2007 Post Count: 120 Status: Offline Project Badges: 
|
Actual runtime was 5 times lower than the estimate for the latest batch of A types. Can this be changed at all? It causes problems.
---------------------------------------- |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Most problems are of own doing with BOINC, but want to hear some gossip :? Was done "accidental" to get a wider spread of the Key A type jobs completed in high priority of which there are only 1000x2 per Target (18 for all of DDDT2), so as to improve chance of being sooner able to make B type batches. By inflating the run times ** your cache filled up real quick :P
----------------------------------------As it stands, my duo fished up 2 and finished both within 24 hours and one is near 72 hours later still waiting on a wingman... that's why ;-) And as by now everyone knows, Type A only goes to "Reliable" devices, so cache should not be big anyhow... tops 2 days. ** Of course this is the same rumor spread as in Alien 2.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
pcwr
Ace Cruncher England Joined: Sep 17, 2005 Post Count: 10903 Status: Offline Project Badges: 
|
I have only ever had 1 DDDT2 WU. Estimated time of 60 hours. Completed in 10.
----------------------------------------Still waiting for more DDDT2 WUs. Patrick  |
||
|
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: 
|
? I thought the reason was because type As used to have 5 times the loops? It was changed a few months ago, but the time for the Wu was never changed.
---------------------------------------- |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The FPOPS time that is stuck in a task header is based of what the feeder has acquired automatically from past work. If the techs don't manipulate that "accidentally", it's stuck until enough is put through the system to auto-adjust that.
----------------------------------------And yes, the A types were to run at 1/5th of what they were at inception, but the C Type would for that run much longer. From the few B Type seen here they would be running longer than A type. So much my recollection ;-)
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
pramo
Veteran Cruncher USA Joined: Dec 14, 2005 Post Count: 703 Status: Offline Project Badges: 
|
but want to hear some gossip When I saw the times go down like they did, that thought honestly was the first thing that crossed my mind  Running a low cache to begin with I managed to pick up 14 of them accross all boxes- all are in and valid except the very first one- And the wingman had a 10 second head start! ts01_ b310_ ps0000_ 0-- - In Progress 11/19/10 20:41:41 11/27/10 20:41:41 0.00 0.0 / 0.0 ts01_ b310_ ps0000_ 1-- 617 Pending Validation 11/19/10 20:41:31 11/20/10 08:40:46 10.08 196.6 / 0.0  |
||
|
|
cw64
Advanced Cruncher Joined: Oct 6, 2007 Post Count: 120 Status: Offline Project Badges:
|
Most problems are of own doing with BOINC, but want to hear some gossip :? Was done "accidental" to get a wider spread of the Key A type jobs completed in high priority of which there are only 1000x2 per Target (18 for all of DDDT2), so as to improve chance of being sooner able to make B type batches. By inflating the run times ** your cache filled up real quick :P My last WU was returned at 22/11/10 18:46:40, so if there are still workunits to be returned ~5.5 hours after this then the net result is not to improve the chance of being able to make B types sooner but is infact decreasing the chance. If you want to increase distribution do it per core/thread, not by messing with my cache tyvm. |
||
|
|
|