| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home Phase 2
Thread: Run dry already? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 137
|
|
| Author |
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
As yet I've not received a single WU as far as I can tell. The chances of getting a FAHB task [lowest feeder priority] is very small when having also selected a second WCG project [WEO CEP2]. In effect, the few feeder slots are continuously vacuumed by those who only do FAHB. What to do (What I do): Got OET+CEP2+FAHB for laptop profile. Now and then switch the profile to only catch FAHB for a day, then switch back to OET+CEP2. Set app_config to only do 2 FAHB at a time when 6 are queued. As the machine needs 24 hours to complete 2 FAHB, these 6 are good for 72 hours continuous computing. When down to 2, fetch 4 again and on and on. Maybe the techs would want to consider the same mechanics as for CEP2, buffer a max number of FAHB. My 8 core laptop anyway starts buckling when running more than 2 FAHB... somehow it causes CEP2 then to too show substantial degrading of efficiency down from 90+% to 70-80%. With 7 OET and 1 CEP2 all is fine, near max efficiency. [Oh dear, did I repeat the wine [sic] again ;?] |
||
|
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1320 Status: Offline Project Badges: 
|
SekeRob* wrote:
@deltavee FAH2_ avx17558-ls_ 000076_ 0004_ 001_ wcgfahb00040000_ 0 is the name format of a partial task to complete 100,000 steps. That I'd like to see confirmed, partial, My understanding was this one would package again 100000 steps to go to 140000. ca05065 wrote: Until I get a partial work unit and can look in the stderr file, I do not know which interpretation is correct. I've got 2 of those 'to continue' tasks: FAH2_avx101121_000083_0043_001_wcgfahb00030000_0 FAH2_avx101121_000033_0046_001_wcgfahb00060000_0 Although at 11% and 8% progress now, it's very likely, that they will behave like the BETA's. Continue with step 30000 to 130000 and step 60000 to 160000. In the stderr: INFO: Checkpointed. Progress x thousand of 100000 steps complete. I will be able to complete them within the 4 days deadline, so I'm able to follow the 9 trickles and finish. |
||
|
|
Byteball_730a2960
Senior Cruncher Joined: Oct 29, 2010 Post Count: 318 Status: Offline Project Badges: 
|
Getting work at the moment....although it has been about timing today with some dry spells during the day.
|
||
|
|
Localizer
Cruncher Joined: Feb 10, 2006 Post Count: 12 Status: Offline Project Badges: 
|
........... Work seems to be flowing.
|
||
|
|
Zarozinia
Senior Cruncher England Joined: Aug 28, 2005 Post Count: 226 Status: Offline Project Badges: 
|
Thank you Seke switching projects around seems to have worked.
---------------------------------------- |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 1950 Status: Offline Project Badges:
|
Seems the "dry spell" is over now and green pastures are in fron to f us, I seem to get almost exclusively FAH2 WUs on pretty much all my hosts...
---------------------------------------- |
||
|
|
ca05065
Senior Cruncher Joined: Dec 4, 2007 Post Count: 325 Status: Offline Project Badges: 
|
I have looked at the output in the Results Status for work unit FAH2_ avx101119_ 000001_ 0044_ 001_ wcgfahb00060000_ 0.
It did do 100000 calculation steps and not the 40000 I expected for this _001_ job. What is the naming convention for the follow on unit? Is it of the format _002_ or does it include a wcgfahb00060000_ suffix? |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
Hello everyone,
The dry spell should be over, I have increased the weight of the project, so we should start to see an up turn in runtime per day on the project as well. As for questions on workunits like what ca05065 has seen, Every work unit sent out has the command to run 100k steps. The researchers have said that they would like to have 100k steps per generation. Going over is good as it provides more details. So in the example provided by ca05065, the first generation _001 did a total of 160k steps (just more details for the researchers). As for the next generation, it will start with just _002. It will not have the wcgfahb00060000. From a quick look at results returned, less than 5% of the results are not returning 100k steps. This is great, because that means 95% of the results coming back go on to generate the next generation. I have seen some 7th generations returned already. Thanks, -Uplinger |
||
|
|
Dayle Diamond
Senior Cruncher Joined: Jan 31, 2013 Post Count: 452 Status: Offline Project Badges:
|
Very cool, thanks for the info Uplinger!
Do we have any word on how many generations the scientists want? |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges:
|
Very cool, thanks for the info Uplinger! Do we have any word on how many generations the scientists want? At the moment the batches with -ls are set for 40 generations. With each generation for the workunits with -ls, they have different parameter changes. The batches with out -ls have 30 generations and no parameter changes. Even though these are the numbers given to us now, we expect the generations to possibly change with future batches. Thanks, -Uplinger |
||
|
|
|