|
21)
Message boards :
Number crunching :
Damn wingman
(Message 24351)
Posted 14 Jul 2012 by  Igor Zacharov
Post:
The redundancy will be needed in the future also, since we cannot exclude faulty devices (not talking about cheating). In particular, one of the side effects of a large scale accelerator study is singling out hosts that produce wrong results. We have seen an indication of that in the past, when getting results from Overclokers, but did not do a systematic study of these effects. With the latest executable this is within reach. Just one explanation. The sixtrack program is for the accelerator study. It needs bit-accurate reproducibility to accurately frame out the appreture. If the results have artificial scatter we cannot zoom to the desired phase space boundary, even collecting order of magnitude more statistics. Eric McIntosh should give more explanation on this and the impact bit-reproducibility will have in science. |
|
22)
Message boards :
Number crunching :
No Tasks ???
(Message 24350)
Posted 14 Jul 2012 by Igor Zacharov
Post: interesting. There should be no difference between linux and windows for the #tasks downloaded. Of course, there is an influence of the past results. So, if the linux boxes were much slower, their average turnaround parameter is higher and they are considered "less reliable" by the system and therefore get less work. This should correct itself over time, as new results come in.
|
|
23)
Message boards :
Number crunching :
How is the SSE3 thing coming along?
(Message 24347)
Posted 14 Jul 2012 by Igor Zacharov
Post: The uptake of the optimized executables (sseX, etc.) is understood and quite stable now. Unfortunately, we had to retract darwin exec, because it was crashing when linked with the latest boinc zip library. Apple will take more time to understand and to correct. Concerning the timeout, apparently there was a much too optimistic factor estimating performance of the sseX, which lead to incorrect projection for the time to run the program. This has been hopefully corrected with the latest recompile of the boinc scheduler at about 7 pm CET on friday. We will monitor this, but at least the tests with our machines tonight show an improvement. Please, report if you see errors your end. The relative performance of sixtrack runs can be compared only with identical input, since simulated particles can hit the wall and terminate the program before they reache 1M turns. Igor. |
|
24)
Message boards :
Number crunching :
Failed to download
(Message 24345)
Posted 14 Jul 2012 by Igor Zacharov
Post: When downloading tasks the scheduler is solving a complicated equation of limits and past execution results. As we had may errors provoked by the server on our side, hosts became "unreliable" and would not download tasks, especially which are not new. This will hopefully clear up in the next few days, as new work is submitted and executed with good results. We sincerely hope for this on our side. Igor. |
|
25)
Message boards :
Number crunching :
Damn wingman
(Message 24343)
Posted 14 Jul 2012 by Igor Zacharov
Post: The mechanism for redundancy is there. We always had the flag <one_result_per_user_per_wu/> in the config.xml file. It seems that we get the dublicates send to the same user/machine when we also enable the matchmaker schedule. It shouldn't happen in my opinion, but this is the fact. We have switched on the matchmaker scheduler, because with cache-job only it would is picky about the hosts it would send work to. We don't need the homogenious redundancy, it fact, we want to compare all possible combinations of computers to study the reproducibility of the sixtrack program. But they indeed should all be different computers. Anyway, in short, there is more experimentation to do with the job distribution algorithm we use for LHC@HOME. I will switch the matchmaker off now. Let's see if we get any dublicates sent to the same user during saturday/sunday. Igor. |
|
26)
Message boards :
Number crunching :
Server aborted my task
(Message 24287)
Posted 11 Jul 2012 by Igor Zacharov
Post: We had to do some drastic changes in order to clean up the data base. As the result of disk running full the data base became inconsistent. This is the reason for download and upload errors over the past 5 (or so days). After trying to correct some of it by hand, we decided that it is better to cancel all outstanding work units. Most of them have been processed anyway already and credit assigned. I apologize to people who contributed the computer time but did not get the due credit for it. Please, write me a private message such that we can compensate the credits appropriately. Thank you, Igor. |
|
27)
Message boards :
Number crunching :
who do I contact about security matter?
(Message 24286)
Posted 11 Jul 2012 by Igor Zacharov
Post: Mike, typically the teams are imported from all other boinc sites and again distributed to all other boinc sites. We get repeatedly complains about teams moving ownership We can change the local data base and hope these changes are eventually propagated outside. Igor. |
|
28)
Message boards :
Number crunching :
Mac OS X Errors
(Message 24044)
Posted 3 Jul 2012 by Igor Zacharov
Post: thank you for the warning. The MAC version was indeed intended for Apple machines with Intel processors. I think we need to analyze this one to see why it is crashing. Sorry about the difficulties. Igor. |
|
29)
Message boards :
Number crunching :
Linux vs. Windows app
(Message 24034)
Posted 3 Jul 2012 by Igor Zacharov
Post: All, we have prepared some optimized sixtrack executables. Hopefully, the machines will now be selected correctly based on the available feature set (sse3). Will do more of these in the next days (sse2 is coming). There is also a MAC executable (10.5 and up, let us know if you need others). If you notice problems, please send us a note. I apologize in advance if we have missed any machines or some platforms do not run. This should be rectified if you send us the warnings. Thank you. |
|
30)
Message boards :
Number crunching :
Two diffrent executables ?
(Message 23961)
Posted 8 Jun 2012 by Igor Zacharov
Post: they are exactly the same. Just some mishap when entering new versions created several files. Hi, |
|
31)
Message boards :
News :
Sixtrack server migration today
(Message 23955)
Posted 7 Jun 2012 by Igor Zacharov
Post: Conan, you are probably hitting on the old address in the cache which points to the old machine. That machine is indeed shut down for maintenance. Try to flush out your cache. Igor. |
|
32)
Message boards :
News :
Sixtrack server migration today
(Message 23935)
Posted 6 Jun 2012 by Igor Zacharov
Post: Friends, last year some experts downloaded by hand our optimized executable for Linux, which would run faster on their specific machine. We could not make it available universally, because it had special instructions not available on older machines. Now, all the runs with that old executable do not pass the validation step, because processing has changed. Please, remove this old executable, such that new one can be downloaded by boinc. We do pay attention to optimization. The new executable should contain code for different machines. If for some machines it doesn't work, let us know. Thank you, Igor. |
|
33)
Message boards :
News :
Sixtrack server migration today
(Message 23927)
Posted 5 Jun 2012 by Igor Zacharov
Post: Friends, we have updated all elements in the tools chain. This includes the machine, as well as the boinc server version and especially the sixtrack program. The new program will allow us to have much better accuracy in determining the properties of the accelerator. As a side effect it also will allow to capture any hardware errors in your machines, should these happen. (Overclockers! Be aware!) A major study has been launched to verify our assumptions about the program computational accuracy and properties of hardware. We would like to verify that the program is running on the newest and also the oldest hardware. Please, let us know if you notice any problems. I apologize for any glitches that may have happened with migration of the data base and other mishaps with the system. Obviously, in case you lost points during the migrations we will compensate. Thank you, Igor Zacharov. |
|
34)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23862)
Posted 20 Jan 2012 by Igor Zacharov
Post: I have taken SEIT.USA and Czech National Team -6M. Please, continue watching the statistics. I prefer the system to do the necessary credit calculation automatically. But since we got involved into the manual correction, we may need to do more adjustments based on the emerging accumulation. |
|
35)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23861)
Posted 20 Jan 2012 by Igor Zacharov
Post: ... It is a valid question that must be answered. The credits are assigned based on the average of the claims from two valid results. Therefore, if one of the valid results claims disproportionate credit, the granted credit for both is still too high (for example (0+1Million)/2=500K). Based on the assignment logs I have, I substracted the unreasonable claims even from these crunches, who where not actively involved in presenting the unreasonable claims, but got inadvertently assigned high values due to the average calculation. After two episodes of this game, we started to calculate the credits based on the internal characteristics of the application. This prevents all cheating and also filters out random/faulty computer clock readings. I hope this explains. I appreciate the work you do. Please, keep crunching. |
|
36)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23858)
Posted 20 Jan 2012 by Igor Zacharov
Post: I have adjusted: Pireatenpartei Deutchland -25M (put it at 150K) Overclock.net -2M Boinc Synergy -2M Final Front Ear -2M Canada -2M Belarus United -6.5M (put it at 150K) However, Seti.USA (12.8M) and Czhech National Team (15.7M), should be both corrected down by 6M based on the statistics provided. I'm not sure of the consistency of the data base, since the present scores seem to be quite high and I don't understand them. Of course, if there are no other ideas, I will just go ahead with -6M. Just wanted to ask first. |
|
37)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23856)
Posted 20 Jan 2012 by Igor Zacharov
Post: Should I assume that team total_credit should be equal to the sum of the total_credit of all the team members? |
|
38)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23849)
Posted 17 Jan 2012 by Igor Zacharov
Post: I have to correct myself here - it was Stelf who was helping with testing. Apparently, on that computer the elapsed time is ok, but the cpu time is bogus. As a wild guess, it could be a problem of units (ticks, or frequency wrong). At any rate, we don't mind, but Stelf - you may want to look at it at some stage. Thank you. interestingly, on 14th of january there was another attempt at the credit sys |
|
39)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23846)
Posted 16 Jan 2012 by Igor Zacharov
Post: interestingly, on 14th of january there was another attempt at the credit sys from Mr. badgoes. I guess, it is just to test if the system is solid enought. Thank you for trying. The system is solid. |
|
40)
Message boards :
Number crunching :
Faulty Computers or Modified BOINC ?? Huge Credits
(Message 23845)
Posted 16 Jan 2012 by Igor Zacharov
Post: I have now adjusted the total_credit entries in the sixtrack data base for wdsmia - 500K Galeon 7 - 400K MrOctane - 900K Jnargus - 600K Sorry guys. Keep on crunching and the scores will come! |
Previous 20 · Next 20
©2024 CERN