| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Are finished WUs pushed or pulled? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 16
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Was there not supposed to be the new policy (in new client of course), that would postpone work requests until uploads were completed? Can't remember if it is in 6.12 or in 7. The fun bit, I've commented on that before. if one runs CEP2 with a zero connect, zero cache. You'd be facing 5-10 minutes of upload time CPU idling. Hope the developers have taken that into account to override that rule, if it is real.
--//-- |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: 
|
Was there not supposed to be the new policy (in new client of course), that would postpone work requests until uploads were completed? Can't remember if it is in 6.12 or in 7. The fun bit, I've commented on that before. if one runs CEP2 with a zero connect, zero cache. You'd be facing 5-10 minutes of upload time CPU idling. Hope the developers have taken that into account to override that rule, if it is real. It was added to 6.12, but not certain it's working correctly yet in any client... As for long uploads, it's supposed to wait max 5 minutes before asking for work. As for v7, the new rules for work-request will likely surprise some users, especially users with little or no "Connected..." but multi-day "Additional..."...  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: 
|
6.13, and on some projects the server can contact the clients.
The scheduler has totally changed with 6.13, so standard recommendations across most Boinc projects will have to change, assuming the same scheduler will be introduced with B7 (Connect about every x days will need to be higher than the Additional work buffer). |
||
|
|
mfbabb2
Senior Cruncher USA Joined: Feb 18, 2011 Post Count: 361 Status: Offline Project Badges: 
|
To send results immediately:
----------------------------------------Add to (or create) cc_config.xml in Boinc Data directory/folder (the one that has the slots and projects directories/folders: <cc_config>
Murphy
---------------------------------------- [Edit 3 times, last edit by mfbabb2 at Oct 29, 2011 6:57:14 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Plz do *not* advertise this code [rather remove, you still can delete] and if you got this from the manual you know why!
If 1, each job will be reported to the project server as soon as it's finished, with an inbuild 60 second delay from completion of result upload. (normally it's deferred for up to a day, so that several jobs can be reported in one request). Using this option increases the load on project servers, and should generally be avoided. It's intended to be used on computer whose disks are reformatted every day. thx in advance --//-- |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
6.13, and on some projects the server can contact the clients. The scheduler has totally changed with 6.13, so standard recommendations across most Boinc projects will have to change, assuming the same scheduler will be introduced with B7 (Connect about every x days will need to be higher than the Additional work buffer). It doesn't need to be larger, but how useful a large "Additional..." is depends on the project. Example, here at WCG you'll max get 15 tasks in a single request. If you've also running HCC exclusively on a quad-core, and let's say each task takes 2 hours to run, 15 tasks is max 8 hours/core, meaning any "Additional..."-setting larger than 8 hours won't give you any more HCC-work since BOINC-client only asks for more work if cached work drops below "Connected...". If on the other hand the same quad runs CEP2 exclusively and is so slow it's hitting the 12-hour-limit, 16 tasks would be 2 days/core, so an "Additional..." below 2 days will partially control the size of the cache. BTW, one of the reasons for changing the behaviour is to decrease how often client connects scheduling-server, since as long as users uses a non-zero "Additional...", they'll get N tasks in a single request, and as long as computer is fast enough to finish N tasks in less than 24 hours, the client will report N tasks on next scheduler-request. (if assumes same-size tasks, something that isn't the standard here at WCG if runs multiple sub-projects). If again assumes WCG-only, in older clients can assume on average 1 task/connection, while with v7 atleast the fastest crunchers will have 15 tasks/connection, something that will reduce the server-load. For any projects that "needs" faster turnaround-time than this, BOINC now includes the option for project to set a per-wu-flag to accomplish this. "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
|