1) Message boards : Number crunching : How can I....... (Message 20951)
Posted 11 Jan 2009 by Toby
If you are getting a "last attempt to recent" error message, it means you are manually pushing the "Update" button before the 15 minute timer has expired. LHC has their server set up so that you can only request work units once every 15 minutes. And on each request it will give you at most 2 work units at a time. Then you have to wait 15 minutes and (maybe) get 2 more.
2) Message boards : Number crunching : Is it possible to inform us.. (Message 20786)
Posted 18 Nov 2008 by Toby
Yes, things are already very limited for this reason. Each time your computer contacts the server you will get at most two work units. Then you have to wait at least 15 minutes before you can request two more. And as was mentioned earlier, each computer can only get 10 work units per core per day. This should do a fairly good job of spreading work out but at the end of the day, some people will get the short end of the stick.
3) Message boards : Number crunching : Segmentation violation (Message 20655)
Posted 13 Oct 2008 by Toby
I just came in to work to find my boinc had crashed again just after midnight on October 12th. It looks like the log may have caught it in the act this time:

12-Oct-2008 00:11:21 [lhcathome] Sending scheduler request: To fetch work. Requesting 84377 seconds of work, reporting 0 completed tasks
12-Oct-2008 00:11:26 [lhcathome] Scheduler request succeeded: got 0 new tasks
12-Oct-2008 00:21:48 [lhcathome] Starting pì~E
SIGSEGV: segmentation violation
Stack trace (12 frames):


It seems to have died while printing the line I hilighted in red. The three characters after the p are actually a two byte sequence of gibberish (when viewed in a hex editor: EC and 85) and I'm wondering if the p is actually gibberish as well... usually that line prints the name of the work unit being started, like this:

11-Oct-2008 23:29:04 [lhcathome] Starting wl55fs_l55f__3__64.3175_59.3275__20_22__6__70_1_sixvf_boinc416043_3

but I don't think any of the projects I'm attached to on this machine have work unit names that start with a p...
4) Message boards : Number crunching : visualisation? (Message 20576)
Posted 1 Oct 2008 by Toby
Unfortunately the LHC screensaver doesn't really have anything to do with the analysis your computer is performing so the answer to your question is "no"

I imagine it would just be a particle going around a big circle... unless the work unit errors out in which case I would expect a giant explosion :)
5) Message boards : Number crunching : Segmentation violation (Message 20575)
Posted 1 Oct 2008 by Toby
I have seen this as well. I didn't really suspect LHC but now that you mention it, I think it has only happened while LHC has had work... but I won't swear to that. I've seen it on gentoo with BOINC 5.10. BOINC is installed through portage so it was compiled on the box it is running on. I even tried reinstalling (which means recompiling) in case some underlying library changed but that didn't help. But it is pretty infrequent which of course means hard to troubleshoot...
6) Message boards : Number crunching : First beam - first real wu\\\'s? (Message 20062)
Posted 9 Sep 2008 by Toby
The project will continue but there are no plans to use BOINC to process actual data at this point.

See the front page news and the thread titled \"What happens to us when LHC goes live?\" for details.
7) Message boards : Number crunching : Statistics Export (Message 19882)
Posted 7 Aug 2008 by Toby
Looks like the names of the XML files changed (from team_id.gz to team.gz and same for host and user). I just updated my database to reflect this and it seems to be working again.
8) Message boards : Number crunching : Not working (Message 19818)
Posted 21 Jul 2008 by Toby
Well the macs will never get work since LHC doesn't have a mac client. As for the PCs... yeah there hasn't been much work in recent memory.
9) Message boards : Number crunching : Ridding computer of Pending Units (Message 19475)
Posted 18 Apr 2008 by Toby
Any chance you could just do a manual clearing of the ancient results? I'm guessing the data is long gone... not sure if it would be as simple as a "delete from result where create_time < 1136091600" (January 1, 2006) but if it is, we could all finally get started on cleaning up our host lists! :D
10) Message boards : Number crunching : Ridding computer of Pending Units (Message 19436)
Posted 14 Apr 2008 by Toby
No :(

Until the admins upgrade the server software and/or manually delete all these records we are stuck with them. Look around these message boards - there are several threads about this topic.
11) Message boards : Number crunching : Waiting On Credit (Message 19080)
Posted 24 Feb 2008 by Toby
Please see one of the other threads that is already discussing this issue like "Stil a pending credit" or "Moldy results"
12) Message boards : Number crunching : Oldest molding/untouched credit awards (Message 19079)
Posted 24 Feb 2008 by Toby
My oldest: 694057 and 23518 from May 3, 2005. That computer is actually still running. It also has 1774356 and 1863985 from July 30, 2005.

Then there is this host of which the hardware still exists but it has been completely reinstalled and is now in use by someone else in the office. It has results 1641366 and 1714116 from July 2005 plus 5 more in 2005.

Host 24201 has 3 results, all from September/November of 2005. This is my old laptop which is laying in pieces in my living room. The memory slots on the motherboard went bad about a year ago so I took it apart to see what was inside.

Host 66594 is one of our two redundant database servers at work. It has 6 results from October/November of 2005 when we were doing some burn in testing before we put them into production in December of that year.

Well those are the hilights at least. Now it is past my bed time :/
13) Message boards : Number crunching : Old inactive Hosts (Message 19078)
Posted 24 Feb 2008 by Toby
I would love to delete my inactive hosts but unfortunately they still have dead results from 2005 attached to them so I can't. And while it would be nice, in the end it doesn't really make that much difference to the project as far as the database goes. The difference between 140,000 (total) and 30,000 (active) host records in the database is really negligible. Seti has over a million in their database...
14) Message boards : Number crunching : granting less credit than claimed? (Message 19040)
Posted 21 Feb 2008 by Toby
Look at what the other computers that worked on the same unit claimed. The granted credit should be some kind of average. Not sure if this is how LHC is doing it but seti used to drop the highest and lowest claim and then average the rest. This is to prevent cheating. Not to say that you are cheating but there are ways of getting your computer to claim WAY too much credit. With this system such ridiculous claims get thrown out.
15) Message boards : Number crunching : Stil a pending credit (Message 18978)
Posted 8 Feb 2008 by Toby
The other problem these dangling results cause is that you can't delete old hosts. I have 16 hosts listed on my account but only 6 of them are active. One of them is my old laptop before I put Linux on it. Now it is actually laying in pieces in my living room because the motherboard gave out a year ago. I would like to forget the dark days of using Windows and move forward to get some closure. But alas I cannot because that host still has old results attached to it on LHC :(
16) Message boards : Number crunching : Boinc farms. (Message 18861)
Posted 21 Jan 2008 by Toby
@Albert: Be very picky about your motherboard! I built a Phenom system in early December and I still don't have it stable under load. It boots up and idles just fine but if I load all 4 cores with seti@home work, it spontaneously reboots after a minute or two. LHC@home tends to last longer - probably because it doesn't hit the memory as hard. But If I restrict BOINC to 2 cores, I can run all day long. Based on various forums and reviews it looks like there are memory compatibility problems. I have a Gigabyte GA-MA790FX-DS5 board with 2 GB of Kingston HyperX 1066 MHz memory. Apparently the phenom and/or the 790FX chipset does not play well with 1066 MHz memory.

One problem is that the Phenom wants to run the memory at 1.8V but almost all of the 1066 MHz memory out there requires 2.2V or more. Most boards let you overvolt the memory and achieve the correct voltage but there have been reports that setting it to over 2.2V can fry the Phenom's memory controller.

I have also had problems with the BIOS just ignoring some of the settings I change. For example, I currently have the BIOS set to put the memory controllers on the Phenom into "ganged" mode but both the POST screen and the AMD OverDrive utility says they are in un-ganged mode. Memory timings are also ignored sometimes. This is probably a Gigabyte BIOS problem.

I just ordered some 800 MHz memory that is on Gigabyte's official "approved memory list" for this board so I'll see if that fixes anything. I think I defintely paid an early adopter tax this time :(
17) Message boards : Number crunching : Info on screen saver? (Message 18464)
Posted 30 Oct 2007 by Toby
Yes, that is correct. It is just simulating some particles interacting with each other in zero gravity. There was some talk a long time ago about making the screensaver display the actual calculations but I think that fell by the wayside. But at least it is pretty!
18) Message boards : Number crunching : Any Mac work coming? (Message 18251)
Posted 17 Oct 2007 by Toby
Originally there was no mac client because the compiler that they used to make the client did not have a mac version. If there is one now, it would likely only be for the newer intel macs. Pretty sure PPC will never be supported on LHC - at least for the sixtrack application that we are currently running. Maybe they will make a mac version of the new Garfield client though...
19) Message boards : Number crunching : Question about CPU (Message 18250)
Posted 17 Oct 2007 by Toby
As I recall, the problem with running BOINC on a cluster is that the core client uses shared memory to communicate with the science application and this does not work across nodes on a cluster. So yeah, you would have to have a BOINC instance running on each node, even if you come up with some way of merging them into looking like a single computer to the project(s).
20) Message boards : Number crunching : Maximum daily WU quota per CPU? (Message 18174)
Posted 15 Oct 2007 by Toby
Or (once again, IF you have permission) you could bring the .exe in on a USB stick with an app_info.xml file. Then it would only have to download the work unit data files which are not .exe files.

Next 20

©2021 CERN