|
1)
Message boards :
Number crunching :
SixTrack and LHC@home status
(Message 23511)
Posted 15 Oct 2011 by Ver Greeneyes Post: I may also have to replace the Fortran Formatted Input/Output for decimal binary conversion by a C routine, using strtod() for the initiated, to maintain 0 ULP difference and precision. Note that Printing floating-point numbers quickly and accurately is hard. But maybe you'll find some inspiration in that paper :) Then at last I might be able to really work on a Floating-Point Error Analysis of SixTrack to determine if we can improve the accuracy of the simulations. Have you considered / are you using Kahan-Babuška-type summation to minimize the accumulation of errors? I recommend this paper for a good understanding of the error reduction / performance trade-off. |
|
2)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 22657)
Posted 5 Jan 2011 by Ver Greeneyes Post: In the meantime I have been busy developing Out of curiosity, does this involve Kahan-Babuška Summation related algorithms (see e.g. http://cage.ugent.be/~klein/papers/floating-point.pdf)? Aside from that, of course it's important to avoid catastrophic cancellation (e.g. use (x+y)*(x-y) instead of x²-y²). |
|
3)
Message boards :
Cafe LHC :
Cyberscience summit
(Message 22425)
Posted 20 Jul 2010 by Ver Greeneyes Post: Wow, good find :) I hope they find a solution for the data quantity problem - bandwidth and HDD storage space aren\'t a problem on my end, but I imagine you\'d need a dedicated serverfarm to service an entire project with the amounts of data cited. |
|
4)
Message boards :
Number crunching :
Don\'t look at the graphics !
(Message 22238)
Posted 17 Apr 2010 by Ver Greeneyes Post: There\'s been quite a bit of discussion about this already, but that obviously won\'t help people who have been attached for a while and run into this for the first time. Personally I wouldn\'t expect much out of the project until they finally update the damn server software (which neither Neasan nor bigmac are responsible for). I think that\'s what\'s really holding everything up, and it would certainly make using the forum a lot nicer. |
|
5)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 22130)
Posted 26 Mar 2010 by Ver Greeneyes Post: Any word on the server upgrade from the QMC guys? I imagine you\'re all very busy with the 7TeV collisions coming up on March 30th, but ... |
|
6)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 22094)
Posted 18 Mar 2010 by Ver Greeneyes Post: Probably just a bad initial estimate. BOINC should correct it over time, but considering how often this project actually has work ... (of course, I\'m hopeful that with the coming server upgrades and everything, the scientists will be able to throw some larger quantities of work our way) |
|
7)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 22092)
Posted 18 Mar 2010 by Ver Greeneyes Post: The graphics are currently broken. Please disable the screensaver, if you use it, and do not attempt to open the graphics manually. Hopefully this will be fixed soon, but we might lose graphics altogether for a while. |
|
8)
Message boards :
Number crunching :
LHC is conflicting with other projects
(Message 22089)
Posted 17 Mar 2010 by Ver Greeneyes Post: Well, 6.10.39 is now out, including the latest scheduler fix. You can find it here. Remember that this is an experimental build - though I don\'t anticipate any problems. |
|
9)
Message boards :
Number crunching :
LHC is conflicting with other projects
(Message 22088)
Posted 16 Mar 2010 by Ver Greeneyes Post: I\'m not sure whether the latest fix made it into the latest build - 6.10.37 - the fix is on Branch but the release was never tagged. Ah, to answer my own question, no it wasn\'t. So I guess we\'ll have to wait for 6.10.38 before we can see if the change finally fixes things. |
|
10)
Message boards :
Number crunching :
Maximum elapsed time exceeded
(Message 22085)
Posted 15 Mar 2010 by Ver Greeneyes Post: The time limit is how long a WU should take to process and if it hasn\'t something is broken and it is cancelled. LHC@home WUs are usually short enough so we had ~24hours as the crunching time, however the new work takes longer than that (which we were unaware of at first). Aah, now I see. In that case, perhaps an absolute limit is simply the wrong approach; a better way might be to check on the completion percentage periodically - stop the WU if it hasn\'t gone up at all in a while, or if it hasn\'t gone up more than some reasonable limit. Of course, that only works if the application is multithreaded, if you can tell BOINC to check up on it in that kind of detail, or if you can change the internal loops to check on themselves periodically to avoid deadlock. |
|
11)
Message boards :
Number crunching :
LHC is conflicting with other projects
(Message 22082)
Posted 15 Mar 2010 by Ver Greeneyes Post: Indeed, the scheduling behaviour has been changing a -lot- over the past releases. I believe the 6.10.x builds introduced a strict first-in-first-out scheduling mode, and since then there have been various attempts to fix resulting task preemption issues. I\'m not sure whether the latest fix made it into the latest build - 6.10.37 - the fix is on Branch but the release was never tagged. |
|
12)
Message boards :
Number crunching :
Maximum elapsed time exceeded
(Message 22081)
Posted 15 Mar 2010 by Ver Greeneyes Post: What is this time limit, anyway? The WU deadline? The only reason to decrease that would be ensure that BOINC reports the results more quickly after the WUs have finished - is that really such a big issue? |
|
13)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 22077)
Posted 15 Mar 2010 by Ver Greeneyes Post: The present Ouch - is there any chance you could migrate to MinGW? (perhaps with the version of GCC offered here until the update to GCC 4.5.0 is finalized) If the client graphics get in the way of such a switch, I\'d just tear them out for now until the code becomes more manageable.. I know some will complain, but I don\'t think a lack of volunteers is the problem here ;) |
|
14)
Message boards :
Number crunching :
Long time pending credits
(Message 22057)
Posted 13 Mar 2010 by Ver Greeneyes Post: I would wait for the server update to be completed. I wouldn\'t be surprised if they purge a lot of old stuff in the process, and it should at least make it easier for them to address these issues. |
|
15)
Message boards :
LHC@home Science :
Black hole question
(Message 22037)
Posted 10 Mar 2010 by Ver Greeneyes Post: Our Universe is fundamentally casual (on microlevel). That\'s why physics says \'Who knows?..\' ;) So, what are you talking about? :) Welcome Back2black! ))))) ♀ The whole point of quantum mechanics is that the universe does -not- appear to be fundamentally causal - all attempts at finding a \'hidden structure\' to explain locality problems have so far failed. Indeed, it really does appear that the answer to \'if a tree falls and there\'s no one around to hear, does it still make a sound?\' is no, at least for the fundamental particles. Einstein may not have liked the idea, and there are still theories (of quantum gravity, no less) that state that the universe is fundamentally deterministic, but at this point we have little reason to question the fundamental randomness of reality. Mind you, even a deterministic reality can be unpredictable - this is a consequence of emergent chaos, which we see just about everywhere in nature; for instance, because the orbits of the planets in our solar system are elliptic, and there are more than two bodies involved, they form a chaotic system, and it is almost impossible to accurately predict their orbits more than several millions of years into the future or past (though we do our best using perturbative theories). |
|
16)
Message boards :
Number crunching :
Another shut down
(Message 22036)
Posted 10 Mar 2010 by Ver Greeneyes Post: According to CERN on Twitter, the BBC article simply restates what was already announced in February. In other words, this is not a new setback - it\'s what they decided was the best way forward a month ago. |
|
17)
Message boards :
Cafe LHC :
Why Does This Board Introduce Spurious Characters In Posts?
(Message 22000)
Posted 8 Mar 2010 by Ver Greeneyes Post: It\'s a bug in the (unreleased alpha) server software they\'re using. It should be gone as soon as they update to a newer version. |
|
18)
Message boards :
Number crunching :
Computation error
(Message 21953)
Posted 5 Mar 2010 by Ver Greeneyes Post: Like others, I recommend they just tear it out until they get a chance to fix it. The screensaver is nice, but like this there\'s no point in it. |
|
19)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 21920)
Posted 3 Mar 2010 by Ver Greeneyes Post: I thought it had been updated! and that was why we couldn\'t run anymore. What I can I tell QMC who are managing the server? Ah. Well first of all, don\'t get cross with them at my say-so: I have literally zero experience running a BOINC server, so take my advice with a grain of salt. However, if you check the Server Status page, you will see it lists the Server software version as 15379. This changeset, pushed on the 9th of June 2008, was never marked as stable, and ironically the very next stable release, changeset 15398, was supposed to fix handling of quotes and slashes for the forum software. The most recent version is changeset 20149, and you can find a history of major changes for all changesets on this page. To summarize: the server software you\'re currently running is not a stable release, and is almost two years old. I recommend getting them to update to the latest version, but be aware that there have been major changes since 2008, and the update may not be a simple one (according to the ServerUpdates page I linked to above, two of the updates require a database update, and this process was only automated in changeset 16160, which the current version predates). |
|
20)
Message boards :
Cafe LHC :
Whatever happened to LHC@home....apologies and thanks.
(Message 21908)
Posted 1 Mar 2010 by Ver Greeneyes Post: Awesome, great to hear. With things on the app side getting in motion, is there any chance the server software will be updated? I know you don\'t have GPU apps or anything, but it seems very behind the times at the moment, and I\'m sure there have been some important bug fixes (maybe something to fix the odd backslashes on the forum as well? One can hope..). |
Next 20
©2024 CERN