| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: Network activity of boinc and wcg? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 11
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Is there a way to force WCG to always have > 5 tasks in the line waiting for CPU? My computer only gets two per a session, and when it's done with them, it stands idle until it fetches more tasks... Is there a way to force it to always have at least 5, or preferably 10 tasks waiting in the line for processing?
Thanx |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The default in old device profiles was to fetch no more than 2.4 hours of work (0.1 days). The new members have a default of 0.3 days (7.2 hours). If you go into the Device Manager > Device Profile you could set it to e.g. 1 day.
----------------------------------------If you are online when the last task finishes, or a while before finishing, BOINC should fetch new work. You need to tell us more of the agent version and profile settings to understand why yours is not doing that.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Mar 25, 2007 11:28:11 PM] |
||
|
|
retsof
Former Community Advisor USA Joined: Jul 31, 2005 Post Count: 6824 Status: Offline Project Badges: 
|
The scheduling in BOINC 5.8 is a bit different than 5.4. In any event, the days between network visits in the BOINC profile is what needs to be changed. If you set it too high, the scheduler doesn't do anything until all jobs are done.
----------------------------------------At 1.0 days, I was only getting 6 hours of work. Start there and gradually increase it by 0.5 day steps. 2.5 days seems to be working well, as a new workunit gets downloaded as the queue gets smaller and I have workunits waiting. This is only for WCG. Differences may be apparent if you run more than one BOINC project.
SUPPORT ADVISOR
----------------------------------------Work+GPU i7 8700 12threads School i7 4770 8threads Default+GPU Ryzen 7 3700X 16threads Ryzen 7 3800X 16 threads Ryzen 9 3900X 24threads Home i7 3540M 4threads50% [Edit 2 times, last edit by retsof at Mar 26, 2007 1:15:19 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The only setting I see is Connect to network about every __ days
So.. would a lesser number make me have shorter work units? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello jhtrico1850,
You have found the right parameter. Retsof is correct. The scheduling algorithm is complex but depends on the expected length of the work units and the time limit. It tries to maintain a large margin of safety, so that it does not draw more work units than it can crunch in the time limit. For work units with a 1 week time limit, the maximum cache size is found at about 2.4 to 2.5 days. I suggest that if you want a large cache, start with 1 day and see if you need to increase it. Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, here is my current Default profile. Note that I've updated clients, so they run with these settings. I'll let it for a few days and then report back the behaviour:
Participate in All Projects Do work while computer is in use? YES Do work only after computer is idle for: 1.0 Do work while computer is running on batteries? YES On multiprocessors, at most use: 4 processors Project Weight: 100 Switch between applications every: 120 minutes Use no more than: 4.0 Gigabytes of disk space Leave at least: 1.0 Gigabytes of disk space free Use no more than: 90 % of total disk space Write to disk at most every: 99 seconds Leave applications in memory while preempted? YES Confirm before connecting to Internet? NO Disconnect when done? NO Connect to network about every 4.0 days Maximum upload rate: 30.0 Kbytes/sec Maximum download rate: 30.0 Kbytes/sec I just want to note that I've had 0.3 days, 0.5, 0.8, 1.0, 1.5, 2.0 and 2.5 days and now I've increased it to 4 days to connect network, and now I only have 7 hours of work in my que, just two tasks. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Your Work buffer is set too high, 4 days in your post. If I remember correctly this project has 7 to 9 day deadlines so the maximum queue is accieved by setting your "connect every" preferrence to 2.82 to 3.76 days. I would recommend the closer to the lower setting.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
9 Days it is for all WCG BOINC projects presently. The occasional rush job of 1 or 2 days will throw a curve ball for the duration of those jobs..... those crunchers are not worried anyway as they go to machines with high reliability <=18 hour average return clients.... their buffers cycle anyway at a very high rate. If it were a WCG only client the above quoted buffer sizes are about best. It gets less if faster projects come thru from other DC's!
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OK, I've lowered to 1.1 days.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I had to delete my boinc dir and reinstall it, and now it seems to work. I have almost full 24h of fetched work! :)
|
||
|
|
|