| 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: Ration WU's, Increase Distribution |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 3
|
|
| Author |
|
|
cw64
Advanced Cruncher Joined: Oct 6, 2007 Post Count: 120 Status: Offline Project Badges: 
|
Have you considered rationing the availability of work units? Like with beta tasks, it would help get the processing done faster. I suspect rationing A types would have a big impact on the speed of the project. B types maybe less so.
----------------------------------------e.g A type restricted to 1 task/BOINC/thread B type restricted to 10 tasks/BOINC/thread // No restriction C type no restriction  |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
This is not the first suggestion to do so and practically it's already the case with the A and B types. They go with preset priority:
----------------------------------------A=10 (only to regular under 2 days returners with less than 0.2% error rate. B=1 (not sure if that is still the case) and shorter than normal deadline A=8 days B=7 days Where before the batches were sub-divided in sets of 25, they are now in sets of 10 to improve the statistical chance of having sets quickly complete. Then there's a very good chance that the techs have enough work to send back for the scientists to generate B from A and C from B with some form of regularity. Presently I venture that the scientists are too busy anyhow verifying the TS units, once the quasi production flow of the 18 targets (not sure if that number stands either), I think much of this button hovering will disappear and clients will just backfill and not like it's now, start with a 0.1 day cache and then when the gauge opens, Dhyson vacuum pushing cache up and up and up. Remember there are: 18,000 x 2 = 36,000 A types to do 36,000 x 2 = 72,000 B types to do 72,000 x 250 (per B type) = 18,000,000 as minimum, but some will require more, so est. is 22,000,000. With run times of new A and B types, we now have good indication what the total computing time is that's needed for the 18 targets A/B and soon we'll know what C type will be (longer than the old C type, per uplinger) AND, not in the know if the special 1 in progress per core Beta rule could be set on per-science level. Beta is an entirely separate bin. Would techs answer that question... some things are left unanswered for a reason, the unsaid :D
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Aug 8, 2010 1:48:01 PM] |
||
|
|
nasher
Veteran Cruncher USA Joined: Dec 2, 2005 Post Count: 1422 Status: Offline Project Badges: 
|
its not the fact that jobs are being sucked up by 1 or 2 people or such... its the fact that cause of the nature of this project there are pauses between types of jobs.
----------------------------------------in my opinion it would be nice to release a set of A's... once 80-90% of that A run was complete send another set of A's... rinse and repeat... once they get enough of a B type they can release it too... unfortunately the way it is set up right now the scientists i believe are still proofing the results and thereby need work units fully processed a-c all types to verify everything is working before they release in larger batches. If there were no badges this wouldn't be an issue between alot of people either.. corse badges is one of the reasons people crunch here more. i keep these selected and so far i am almost 1/2 way to my bronze badge here... with the rate of this project i have decided to go to gold currently and then see how things go. With Beta work units they want them tried by many computers but with DDDT2 they want them back so they can get more done. personally i would love to be able to get all badges but i am more concerned with the science being done and us either curing things or progressing things. So i want them done the most efficient way. the scientists are working on making things faster for this project as they go. it just may take some time. hopefully next week or so we should see a new project C4CW .. whatever that ends up to be (currently ending beta) and CEP-2 should soon follow (next month or so) to all projects.. so badge hunters should have a grouping of projects to hunt soon and hopefully when people get to the badge levels they want with those projects or a good level.. by then they will get DDDT2 running faster and we will have that one to crunch as well. Happy Crunching  |
||
|
|
|