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: 14
|
![]() |
Author |
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Greetings,
We are changing the length each work units runs on your machine. The work units in the past have been estimated to run about 10 hours per work unit. This is an average over all the member machines. We are now targeting 6 hours per work unit with the new batches. The new batches that are targeting 6 hours each start with batch 4950. Sekerob has a nice graph in the FAQs that show the runtime averages for each project. http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=18277 The graph I refer to is the labeled "WCG Project Task Average Run Time" Thank you again for your continued participation in World Community Grid! -Uplinger |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi there. I am a bit confused. why are you cutting down on the amount of time need to crunch a fah work unit? won't this stop you from getting the propper data?
as I understand, the long work units are from the most resistant strain of the virus. I am really baffled, can someone please explain? thanks. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The techs are becoming rather good at slicing and dicing AutoDock tasks to average a targeted length. Some applications can do this more easily than others - for example, NRW tasks hit the right time exactly, but CEP tasks are much more variable.
For the exact technical details of how tasks are sliced and diced - I can't recall the techs ever sharing that. Part of it is down to how the researchers specify the work units. Part of it is how they are packaged by the work generator. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello whocrazy,
If you watch the FAAH graphics, you will see that it repeats a number of times and each time it usually produces a lot of temporary dots on the screen. The work unit is really a package that can bundle up more or fewer of these iterations. So the staff can adjust the amount of work required by each work unit. Lawrence |
||
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Hello,
Lawrence and Didactylos are correct. We receive work units from the researchers that are about 30 hours each for these batches. We split these up using a formula on the servers. Shortening the average work unit length does not mean the researchers get less info. It means we send out more copies that are shorter in length to solve all the iterations. Purpose for shortening the length for our members is to allow members to return work more regularly especially members who keep their machine active maybe only an hour per day. This gives their computer a chance to finish the work unit without being considered too late in the result status page. -Uplinger |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
oh wow. I didn't know you could do that.
btw, is it possible to make exceptions for people with fast computers and also for people who keep their machines turned on all the time? IE, the really fast machines like mine could do the 30 hour work units, and the slower ones or the ones that aren't active a lot could do the fast work units. Thanks. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello whocrazy,
This is already on the list. Eventually the server will be reprogrammed to send out long work units to fast computers and short ones to slow computers. I don't know when we will find the time to program this. Lawrence |
||
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1322 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Hello whocrazy, This is already on the list. Eventually the server will be reprogrammed to send out long work units to fast computers and short ones to slow computers. I don't know when we will find the time to program this. Lawrence Hello Lawrence, IMHO WCG had on the list, that we volunteers get the possibility to define in our device profile, whether we prefer long or short Work Units to receive. Cheers |
||
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Crystal Pellet,
The hard part will be getting it to work in the first place. However, once we do, I think that we can add a user preference that influences that size of workunits that you are sent (something to the effect of "on average send me 'longer', 'average', 'shorter' workunits"). However, there is much work to do before we get to that point. thanks, Kevin |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The hard part will be getting it to work in the first place. However, once we do, I think that we can add a user preference that influences that size of workunits that you are sent (something to the effect of "on average send me 'longer', 'average', 'shorter' workunits") My immediate reaction is that I would choose shorter WU to reduce a number of risks I believe are present when running longer WU. What would happen if you ran out of short (or long) WU? Don't you risk the wrath of members if you then send long (or short) WU instead? It doesn't seem to me that users have adequate information in order to make a reasonable decision on the WU length. The length of a WU sent to any computer should be based only on the ability of that computer to crunch the WU. I don't want to see you change the points per hour and base that on WU length. From my view this will raise more questions than it will provide answers, but as you work out the design, maybe the questions will be answered. |
||
|
|
![]() |