Posts by old_user28601

1) Message boards : Number crunching : Odd stuff. (Message 6363) Posted 25 Nov 2004 by old_user28601 Post: > CPDN is very dependent on memory I/O, so even with hyperthreading you're still > hitting the same memory bus etc, i.e. you have four times the I/O which I > imagine is slowing things up. Yes, that could also be. But I gather it doesn't explain the slowdowns The Gas Giant is seeing...? JÃ¶rg
2) Message boards : Number crunching : Odd stuff. (Message 6344) Posted 24 Nov 2004 by old_user28601 Post: I have noticed the CPDN calculations don't behave well under hyperthreading. Initially, I had configured my system to run 4 units in parallel (it is a dual Xeon with hyperthreading). The estimated time for completion (for each of the units was 1600 hours). However, at first, all process did seem to run in parallel (in taskmanager, the cpu time for each unit was equal to real time). But as the model went forward, the parallelism went (in taskmanager, the cputime for each unit was much less than real time). Now, my system is configured to run 2 units in parallel, and this yields an estimated time of 600 hours (for each of the units). This means that if I run two units in parallel, and after that again two units in parallel, I'd have spent 1200 hours to complete. This is a huge difference from the 1600 hours. Apparently, the type of calculations do change over time, yielding less and less benefit from hyperthreading. You can check whether you benefit from it in the taskmanager: just look at the CPU time spent on both workunits. If this is over half of real time, you'll benefit from hyperthreading; if it is less than half of real time, the processes are taking turns on the CPU (no more parallelism, even made worse by the time spent swapping). JÃ¶rg
3) Message boards : Number crunching : large work unit? (Message 6333) Posted 24 Nov 2004 by old_user28601 Post: It was the hyperthreading... Running 2 units in parallel shows an estimate time of 600 hours for the 2 units that are paused. The 2 running units show the same estimated time, but it runs faster. :-) Still, from 1600 hours for 4 units in parallel to 600 hours for 2 units in parallel (thus 1200 hours for all 4 of them) is quite a performance increase. JÃ¶rg
4) Message boards : Number crunching : large work unit? (Message 6230) Posted 19 Nov 2004 by old_user28601 Post: > The estimated time to completion is based on the benchmark scores. > > Have you checked if your "identical" systems have identical scores? They are close. One is a P4 2.4 GHz, the other a dual Xeon 2.4 GHz (hyperthreading enabled). For the dual Xeon, it shows the benchmark scores per logical CPU and they seem to more or less match the P4 benchmarks. The Xeon is doing 4 units in parallel, each with an estimated time of 1600 hours, the elapsed time does change with real time (as does the cpu time spent on the different processes). However, when I move the work from the Xeon to the other computer, it only runs one unit (which is logical) but gives the same time estimate. The Xeon has been running for 160 hours, and the graphs are at 10%, so it is not only the estimate that is wrong: the unit does take longer to work on. > There is a known bug when benchmarking multiple CPU (or HT) systems causing > the benchmark scores to fluctuate greatly. If your computer is HT then this > might be the cause. Yes, but that wouldn't explain why the calculations also take longer...? Would it? JÃ¶rg
5) Message boards : Number crunching : large work unit? (Message 6222) Posted 19 Nov 2004 by old_user28601 Post: Hello, I noticed something strange with the workunit size. I\'m using one computer that doesn\'t have network, and another computer that does. The latter one is also used to download work/upload data for the non-networked computer (I\'m working on a tutorial on how to set this up :)) Now, the workunit on my networked computer is estimated to take 600 hours (over halfway now), but the workunits I downloaded for my offline system are estimated to take 1600 hours. This is even the case when I move those workunits to the networked computer, so it doesn\'t seem to be related to the computer speed (they are both Intel at 2.4 GHz anyway). The main difference in the download was that for downloading the workunits for the offline system, I specified a long time between network access (this default value is 0.1 time per day). Did this cause it? JÃ¶rg
6) Message boards : Number crunching : Result exited with zero status? (Message 6203) Posted 18 Nov 2004 by old_user28601 Post: I'm having the same error on all my machines... :-( - 2004-11-17 12:32:52 [climateprediction.net] Result 3ev9_100180838_0 exited with zero status but no 'finished' file 2004-11-17 12:32:52 [climateprediction.net] If this happens repeatedly you may need to reset the project. 2004-11-17 12:32:52 [climateprediction.net] Restarting result 3ev9_100180838_0 using hadsm3 version 4.04 2004-11-17 14:52:43 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi 2004-11-17 14:52:46 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded 2004-11-17 17:26:49 [climateprediction.net] Result 3ev9_100180838_0 exited with zero status but no 'finished' file 2004-11-17 17:26:49 [climateprediction.net] If this happens repeatedly you may need to reset the project. 2004-11-17 17:26:49 [climateprediction.net] Restarting result 3ev9_100180838_0 using hadsm3 version 4.04 2004-11-17 22:50:15 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi 2004-11-17 22:50:19 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded 2004-11-18 06:21:31 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi 2004-11-18 06:21:34 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded - This is the log from my first machine. The error has occured repeatedly (4 times) in the last couple of days. My second machine is running 4 units at the same time (dual Xeon with hyperthreading). Yesterday, it gave the same error for all units it is working on (after having spent 140 hours on each unit). On both machines, the stderr.txt is empty... Any suggestions? JÃ¶rg
7) Questions and Answers : Windows : client seems to restart the same calculations (Message 6189) Posted 17 Nov 2004 by old_user28601 Post: > Resetting is a last resort, because it does mean scrapping the whole run. It > is done from the projects tab, and right clicking on Climate Prediction. > > Some people experience this fairly frequently, but it doesn't seem to affect > the calculation so I would ignore it for now. Ok... I would be a shame to have to restart: it has been working on this unit for 274 hours (and approx. halfway). JÃ¶rg
8) Questions and Answers : Windows : exited with zero status but no \'finished\' file (Message 6181) Posted 17 Nov 2004 by old_user28601 Post: Same problem here... 2 days in a row, with the same work: 3ev9_100180838_0 2004-11-17 12:32:52 [climateprediction.net] Result 3ev9_100180838_0 exited with zero status but no 'finished' file 2004-11-17 12:32:52 [climateprediction.net] If this happens repeatedly you may need to reset the project. 2004-11-17 12:32:52 [climateprediction.net] Restarting result 3ev9_100180838_0 using hadsm3 version 4.04 Any solutions? (Win XP, SP2) JÃ¶rg
9) Questions and Answers : Windows : client seems to restart the same calculations (Message 6180) Posted 17 Nov 2004 by old_user28601 Post: Hello, Yesterday, I encountered this error, which resurfaced today: - 2004-11-17 12:32:52 [climateprediction.net] Result 3ev9_100180838_0 exited with zero status but no \'finished\' file 2004-11-17 12:32:52 [climateprediction.net] If this happens repeatedly you may need to reset the project. 2004-11-17 12:32:52 [climateprediction.net] Restarting result 3ev9_100180838_0 using hadsm3 version 4.04 - Does this have to do with the server in Bern being down for maintenance, or is something else up? When I got the error the first time, there was a lot of network traffic (I was downloading some large files). Today however, this wasn\'t the case... Must I reset the project (if so: how? And does it mean my previous calculations are lost?)? Or are there other things I should try first? Thanks! JÃ¶rg
10) Questions and Answers : Wish list : using computers without internet access (Message 6179) Posted 17 Nov 2004 by old_user28601 Post: Everything works ! I'm writing the necessary steps (and the possible issues) to make a tutorial for other people that want to do the same. JÃ¶rg
11) Questions and Answers : Wish list : using computers without internet access (Message 5994) Posted 9 Nov 2004 by old_user28601 Post: > > I think that the server will just record it as a second computer when > you > > attach, identical of course to the first. That should be OK, but bear in > mind > > that I don't know whether anyone has actually done what you are > suggesting, so > > it is possible that you will run into a problem somewhere! > > Only one way to find out... > I'll install the client on the non-networked PC this afternoon, and swap the > boinc-folders as planned (steps 1-6). It does indeed create a second computer. Currently it is downloading work, I'll suspend it once all data has arrived, and verify tonight if my home PC does indeed start crunching. (hopefully, I have my settings for multiprocessor correct, as my machine at home is a dual Xeon, but my machine at work is a P4) JÃ¶rg
12) Questions and Answers : Wish list : using computers without internet access (Message 5992) Posted 9 Nov 2004 by old_user28601 Post: > I think that the server will just record it as a second computer when you > attach, identical of course to the first. That should be OK, but bear in mind > that I don't know whether anyone has actually done what you are suggesting, so > it is possible that you will run into a problem somewhere! Only one way to find out... I'll install the client on the non-networked PC this afternoon, and swap the boinc-folders as planned (steps 1-6). We'll see if and how it works... (I'll report back here) Thanks for the assistance! JÃ¶rg
13) Questions and Answers : Wish list : using computers without internet access (Message 5990) Posted 9 Nov 2004 by old_user28601 Post: > I think this covers it. Remember that if you do use CD for transfer then you > will need to write enable the files each time. And at step 4 you will need to > attach to the project before you can download. Ok! > Trickles are not important except in telling the server how you are getting > along. The 'mega trickles' at the end of phases 1 and 2 contain some data, but > you will not need to connect to the network until the end of the run. Well, I'm trying to pass someone (he just added 4 PCs :-(), so it could be fun to upload trickles every now and then. > There is a bug in the current BOINC client which means that you will need to > disable network access each time you restart BOINC on the machine without > network access. Ok! Euhm, one more thing: will it appear as a second computer to the server (even though it will be up/downloaded with the same computer)? If not, I might have to play around with some settings a bit... JÃ¶rg
14) Questions and Answers : Wish list : using computers without internet access (Message 5987) Posted 9 Nov 2004 by old_user28601 Post: > If the computers are networked together, the simplest solution would be to > install a proxy and use that. > <a> href="http://www.boinc.dk/index.php?page=user_statistics&project=cpdn&userid=13"><img> border="0" height="80" src="http://13.cpdn.sig.boinc.dk?188"></a> > They have not network connection at all (physically, they are in different locations). I've read some other threads more or less related, and will try it like this: 1. install BOINC on the non-networked PC 2. make a backup copy my BOINC folder on the networked PC 3. copy the BOINC folder of 1. to the BOINC folder of the networked PC (using CD / USB storage) 4. download the unit 5. copy the BOINC-folder with unit (to CD / USB storage) 6. restore the backup made in 2 (on the networked PC) 7. copy the folder from 5 to the non-networked PC. After the crunching (can I work with trickles, or must I wait for the whole unit to finish?), I would do the same steps. I expect that in step 4, the data will now be uploaded to the server. JÃ¶rg
15) Questions and Answers : Wish list : using computers without internet access (Message 5967) Posted 8 Nov 2004 by old_user28601 Post: > The BOINC folder you refer to (with the renaming), is that > program filesboincprojectclimateprediction.net Hmm, my slashes seem to have gone. I meant program files_boinc_project_climateprediciton.net (the _ should be backslashes) JÃ¶rg
16) Questions and Answers : Wish list : using computers without internet access (Message 5966) Posted 8 Nov 2004 by old_user28601 Post: > You would certainly need to keep the two separate. But unless you know what > you are doing, and are willing to concentrate on getting it right every time, > you are asking for trouble! You could not run both at once. You would have to > suspend and close the BOINC client that you normally ran on the internet > connected machine, rename the folder, create a new BOINC folder and download a > new WU, move that to the other machine, rename the old BOINC folder back to > its own name, then go through the process again when the other one was ready > to upload. That is missing out several steps, which is why you have to > understand the whole process. I have not heard of anyone doing it but it might > work. Isn't it possible to have two seperate installtions of the BOINC client on one PC? That way, each BOINC would have its own folder, and I could just copy one of these clients to and from the offline machine..? If this won't work, I can imagine the manual process to be cumbersome. The BOINC folder you refer to (with the renaming), is that program files\boinc\project\climateprediction.net ? > An easier way to do much the same thing, though you would not earn any BOINC > credits, is to run CPDN classic on the machine without internet connection. > They both run similar climate models, but you would not run into the > difficulties I outline above. See <a> href="http://www.climateprediction.net/board/viewtopic.php?t=1977">here</a> > for the list of clients - v2.2.29 is the one you would choose for this. Hmm... Someone challenged me to pass his number of credits :-) ... So for this the credits would matter. Still, if it doesn't work with the credits, it is nice to have an alternative. JÃ¶rg
17) Questions and Answers : Wish list : using computers without internet access (Message 5962) Posted 8 Nov 2004 by old_user28601 Post: Do I need to install two boinc clients on the computer that has internet access? (i.e. one for local crunching, and a second one to use only for the transfers regarding the offline computer) JÃ¶rg
18) Questions and Answers : Wish list : using computers without internet access (Message 5947) Posted 7 Nov 2004 by old_user28601 Post: I'm not sure I'll go for it, but I'm just checking my options... Thanks for the info! JÃ¶rg
19) Questions and Answers : Wish list : using computers without internet access (Message 5917) Posted 5 Nov 2004 by old_user28601 Post: Hello, Is it possible to use computer that have no internet connection (not even dialup) to perform calculations? I do have access to a computer with a fast connection which I can use to upload/download data (and move it from there to the offline machine). Thanks! JÃ¶rg