Posts by Keck_Komputers

1) Message boards : Number crunching : Enlarge Window of Boinc manager (Message 867731)
Posted 21 Feb 2009 by Profile Keck_Komputers
Using file-exit will sometimes store the window size.
2) Message boards : Number crunching : Boinc 6.4.4 and Seti@Cuda (Message 841254)
Posted 18 Dec 2008 by Profile Keck_Komputers
I havn't seen any problems with 6.4.5 however I don't have a CUDA card so that could make a difference.
3) Message boards : Number crunching : Difference between Boinc and "World Community Grid Boinc" (Message 838188)
Posted 9 Dec 2008 by Profile Keck_Komputers
They are basically the same thing with a different skin. The only incompatibilities you are likely to run into are different folder locations causing issues.
4) Message boards : Number crunching : Has anyone seen a 'trough' in their data? (Message 831081)
Posted 16 Nov 2008 by Profile Keck_Komputers
This is normally a known type of terestrial interference. However it might be a good idea for another SETI project to look for the lack of signals where they would be expected.
5) Message boards : Technical News : Composite Head (Nov 05 2008) (Message 828521)
Posted 9 Nov 2008 by Profile Keck_Komputers
There is a feature in the BOINC server code to prevent (successful) repeat requests for work within a defined period.

LHC@home uses it set to ~ 15 minutes
Many other projects have it set between 1 - 4 minutes
On SETI it is set at ~ 7 or 9 seconds.

Surely increasing the Communication deferral to e.g. 10 minutes would releive a lot of the load on the servers.

Good point here. I have always thought a good idea to help deal with server congestion would be to automatically scale this defferal based on how busy the server is. I would range it from 1 minute when the server is not dropping any connections up to 4 hours when nothing can get through.
6) Message boards : Number crunching : Long term debt, again (Message 820406)
Posted 19 Oct 2008 by Profile Keck_Komputers
When you are first starting out with unbalanced resource shares the client will download up to a full queue of work from the backup project. The client may also get a full queue of work any time it has to contact the backup project. Once that work is processed you usually will not see that project come up again for a long time unless the main project(s) have problems sending work. So as others have mentioned shorter queues generally work better.
7) Message boards : Number crunching : Pending Credit (Message 815542)
Posted 7 Oct 2008 by Profile Keck_Komputers
Currently the tasks must be varified before credit is granted. Sometimes you are the first one to return a task so you get pending. Other times you are the last one to return a task so the credit is granted almost immediately.
8) Message boards : Number crunching : BOINC Mgr "Report Deadline" needs to sort via 24 hour clock (Message 811150)
Posted 23 Sep 2008 by Profile Keck_Komputers
BOINCmgr always uses ASCII sorts. If you set your OS to 24 hour mode you should get the correct sort order, however even then the date may confuse things from time to time. The client sorts numerically so there is no risk of it crunching the wrong one. However earliest date first mode is gone in version 6 clients, the new scheme usually comes up with the same task needing CPU time so you may not notice the difference.
9) Message boards : Number crunching : could dups cause credit problem? (Message 810500)
Posted 21 Sep 2008 by Profile Keck_Komputers
My guess is that your live install did not have a chance to report the task(s) you had finished before your reboot. If the tasks are not reported they are in limbo and will eventually time out and be resent to other hosts. Tasks do not report immediately after upload to reduce the load on the server.
10) Message boards : SETI@home Science : Very Long Anyalisis of a work unit (Message 808212)
Posted 14 Sep 2008 by Profile Keck_Komputers
You have an Astropulse task. That is a reasonable run time for that type of task.
11) Message boards : Science (non-SETI) : September 10th 2008 - Cern LHC switch on (Message 807199)
Posted 12 Sep 2008 by Profile Keck_Komputers
It looks like they have resolved the issue they had with 6.* BOINC, I now have a LHC WU in my cache, and my BOINC client is 6.2.16.

I guess they are testing before the real thing.

Actually 6.2.18 resolves the issues LHC was having, please upgrade.
12) Message boards : Science (non-SETI) : Speed of Light, and space travel. (Message 805388)
Posted 6 Sep 2008 by Profile Keck_Komputers
Observing a signal generated by the cat is a way of observing the cat.

Relativity may only apply locally. Of course in this case "locally" means about 13 billion light years.
13) Message boards : Number crunching : BOINC 6.2.18 for Windows - not ready for prime time (Message 805376)
Posted 6 Sep 2008 by Profile Keck_Komputers
I had it set to use 95% of the processors which should have let ti run both cores like prior versions of BOINC. I reset it to 100% and closed and re-started and the second core started working again. Doesn't make any sense that it should require the 100% setting as long as it was above 50%.

Just being conservative. The percentage always rounds down unless that would result in 0 CPUs in use.
14) Message boards : Number crunching : Some screenies from 21:10 - 21:50 approx Oslo (GMT +2) Saturday 30th, 2008. (Message 803604)
Posted 31 Aug 2008 by Profile Keck_Komputers
The app runs at several different bandwidths or dispersion levels. Some levels can not be processed in some tasks. So it is probably nothing really unusual, just you have not noticed it before.
15) Message boards : Number crunching : About a hundred off calculatin error's on the X64 host (Message 790824)
Posted 1 Aug 2008 by Profile Keck_Komputers
After my last post you have now a lot of:

Can't get shared memory segment name: shmget() failed

I have no experiences with this message.. hmm.. something new..
But I think this is a problem with your PC or the BOINC client or something other.. ;-)

Two examples of your errors.. and your wingmen have well results..

Have you restarted your computer since upgrading to a version 6 client? That is an error I would expect when the new groups do not have their permissions set correctly. A restart will normally fix it, if not reinstall and reboot and see if that helps.
16) Message boards : Number crunching : If one thing goes sour . . . (Message 784257)
Posted 19 Jul 2008 by Profile Keck_Komputers
I'm surprized you get anything returned successfully. 6.1.0 was a very unstable test version that did not even make it to alpha testing. Currently there is a 6.2.14 release candidate that should be better than that version.
17) Message boards : Number crunching : To relocate BOINC directory - how? (Message 782079)
Posted 13 Jul 2008 by Profile Keck_Komputers

I've never tried it, but exiting BOINC, making a copy of the directories, uninstall & re-install, copy the backup directory to where the new installtion is, restart BOINC & it should continue on from where it left off.

This works. There is a simpler way as well. Copy the BOINC directory to the new location then uninstall and reinstall to the new location. This way you usually keep the host ID as well as all the in progress tasks.
18) Message boards : Number crunching : @ Astropulse Rumor - Reality (Message 764034)
Posted 7 Jun 2008 by Profile Keck_Komputers
Since ASTROPULSE is looking wide-band signals, it isn't helpfull in finding ' ALIEN or other (Narrowband) signals.
As Pappa stated, it looks for 'other' phenomenia, like pulsar's, blackholes, etc.
Would take a gigantic amount off power to 'broadcast in W I D E - BAND , mode though, so that isn't an option.

Or am i terrible wrong, on my assumptions here???

Mostly right I think. If I am thinking correctly about the only ET Astropulse would be able to detect is one in the process of destroying itself in a nuclear war. Of course that lacks the repeatablity to prove it.
19) Message boards : Number crunching : V8-Xeon Server (Message 761632)
Posted 1 Jun 2008 by Profile Keck_Komputers

The higher temperature you noticed on core 0 is a feature commented on by Quad core owners. Both my Quads show the same with Core Temp 0.97.1, but your V8 system may be a little higher in it's difference.

Certain OS processes can only run on CPU 0, core 0. It is a little surprizing that this could make substantial difference on a full bore crunching rig but it certainly explains why a general use computer would see a higher temp there.
20) Message boards : Number crunching : How is average CPU efficiency determined? (Message 759244)
Posted 27 May 2008 by Profile Keck_Komputers
The calculation is CPU time divided by wall time. Most likely you had a misbehaving application that had more than one processing thread.

Next 20

©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.