Message boards : BOINC client : Connect every 0.5 day=2 hours of work
Message board moderation
Author | Message |
---|---|
Send message Joined: 30 Aug 05 Posts: 58 |
Hi, Using CC 5.8.11 on Linux. I'm currently just running Malaria@home and when I request 0.5 day of work, I receive 6 Wus of 30-40 min each (on a HT machine: 30-40 min to crunch 2 Wus) The estimated time is around 2h30/wu but the real CPU time is around 30 min. The estimated time doesn't seem to adjust to the real lengh of the Wus when they're finished.. More than that: Boinc doesn't request new work when only two wus are remaining and the network activity is working.....:o( (what's happening if I suspend the network activity at that moment: the machine will stay out of work for 12 hours:it's not normal) Couldn't the Boinc scheduler adapt itself to the real CPU time of the latest Wus and the number of remaining Wus when requesting work because just receiving 2 hours of work when requesting 12 is a little desappointing. Or perhaps, it's Malaria that is restricting the number of wus downloaded...... Edit: Well, the cache is now empty (all the wus are finished) and Boinc isn't even requesting new work.....Sometimes I wonder if you do test before releasing new versions... |
Send message Joined: 30 Aug 05 Posts: 58 |
Ok. I had to stop and restart the service to download new work. It looks like Boinc doesn't like instable internet connection:it was completly freezed.... The DCF seems to work now: the wus are estimated at 1h50 instead of 2h30 but still not enough work to feed my hungry machine |
![]() Send message Joined: 29 Aug 05 Posts: 15598 ![]() |
The estimated time of any result is given out by the project, it's not built in into Boinc. So you can only teach Boinc to get more work, by getting more work, by which the DCF goes down... |
Send message Joined: 30 Aug 05 Posts: 58 |
Ok thanks for the answer (I'm so accustomed to run CPDN wus that I had never see Boinc learning...) And it looks like the first time I attach to Malaria, I received a bunch of short Wus: no luck :o) Now the wus are longer (around 1 hour each) and the estimated time is around 1h40. |
Send message Joined: 30 Aug 05 Posts: 58 |
But I have still a problem with my unstable internet connection: When I update the project, Boinc seems stuck: Sat 17 Feb 2007 09:35:20 CET|malariacontrol.net beta|Sending scheduler request: Requested by user Sat 17 Feb 2007 09:35:20 CET|malariacontrol.net beta|Reporting 2 tasks ................ and nothing more: half a hour later, boinc hasn't tried to communicate again but it's uploading work correctly. In the project tab: 'Scheduler request pending: requested by user' is displayed but there is no time (is it new?). When I press the update button, nothing happens: no message in the message tab to say the server is down but google is working, etc. I have also another problem: when I copy/paste the messages of the message tab of the remote machine: the boincManager freezes or the messages are not copied (I don't see them in Klipper in KDE). When copy/paste on localhost it's working OK.... (and yes, I know I'm a pain...:o)) |
Copyright © 2025 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.