| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Official Messages
Forum: News
Thread: 2023-01-25 Update (ARP & OPN1 workunits) |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 35
|
|
| Author |
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges: 
|
Releasing in smaller quantities would mean the bigger machines scooping more of the units so less for the smaller machines. However, probably fewer "No Reply" . Folks just should stop messing around with the settings for the number of WUs being requested....Mike Ralf Doesn't matter when BOINC / server just ignores user settings anyway. I really wonder why there's no 1 task per thread assignment limit on ARP. It just sounds so obvious... more fairly distributes tasks among users, decreases the chances of getting too many tasks a PC can't finish in time (I'm looking at you, first installation machines) and it reduces issues with computers going idle because files aren't downloaded in the right order. Am I missing something? Well, basically, there is/was. But some folks figured out to work around the default settings and manipulate/game the system for their own selfish reasons...Ralf |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges: 
|
I have 2 machines both with 8 threads. I have a cache limit set in profile of 12 ARP1 per machine and app_config set to a maximum of 4 threads. When a batch is released I receive a maximum of 8 units each but that increases gradually to 12 if there are enough available. It is enough to keep a steady flow without hogging the queue and keeps the speed of processing up.
My machines crunch units in 21-24 hours, so I set the cache to 12 as they would then all clear in 3 days which is half the standard deadline or within the enhanced deadline. My machines operate 24/7. They also have cache settings of 8 each for the other projects with a limit of 4 threads each for them. That keeps all threads operational and allows for 1 project to be out of units. If 2 are down then I have to adjust the survivor to 8 threads. Mike |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7844 Status: Offline Project Badges:
|
Since there has been a nice steady supply of workunits I pared my queues back from 3 days to 2 days. I have had no problem keeping the them filled and having everything usually back in around 2 days. With the two different types of MCM units, one taking about twice as long on Linux as the other one, an overabundance of one over the other will cause some seesawing of the number in the queue but this way there is still plenty of slack to get everything returned with plenty of time to spare.
----------------------------------------I have noticed a number of "Server Aborted" tasks where someone was just over the time limit by a couple of hours, but did get theirs completed before my systems got a chance to get them started. At least there was no time lost in redundant crunching. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2494 Status: Offline Project Badges: 
|
ARP & OPN1 workunits <Snipped rest of Cyclops post>On Monday afternoon, many volunteers reported receiving new ARP1 and OPN1 workunits. These workunits are not from a new batch; these are older WUs that were never sent out due to an overloaded server causing problems in our workunit-distribution process. ARP1 and OPN1/OPNG teams remain on temporary pause, preparing new workunits. WCG team I think we finally run out of these "old" OPN1 tasks, and we now have to wait for new tasks for the new target. Only resends left now, it seems. [Edit 1 times, last edit by Grumpy Swede at Feb 5, 2023 1:07:17 AM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7844 Status: Offline Project Badges:
|
I think we finally run out of these "old" OPN1 tasks Nope. Just got this one OPN1_0128400_2503_1. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges:
|
They are still trickling through - re-sends only.
Mike |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7844 Status: Offline Project Badges:
|
They are still trickling through - re-sends only. Mike Rare but not a resend:OPN1_0127746_00800_1 Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges:
|
Are you sure? OPN1 are usually sent out as singletons, so _1 would indicate re-send.
Mike |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7844 Status: Offline Project Badges:
|
Are you sure? OPN1 are usually sent out as singletons, so _1 would indicate re-send. Mike It is the second unit but it lists quorum 2. Currently pending validation. You may be correct as I had the last of _0 units on Feb 4, 2023. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 442 Status: Offline Project Badges:
|
I have seen mention of EXTREME. What other levels are there in WU and how do I recognize which is which?
Currently running all systems with zero additional queue - I got fed up with server aborted WU that seemed to sit around not doing anything (NO REPLY) for 6-7 days then showing up with minimal processing time causing my work to be aborted. Bruce |
||
|
|
|