Message boards :
Number crunching :
People with bugs, what are your boinc settings?
Message board moderation
Author | Message |
---|---|
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
Do work while computer is running on batteries? (matters only for portable computers) yes Do work while computer is in use? yes Do work only between the hours of (no restriction) Leave applications in memory while preempted? (suspended applications will consume swap space if 'yes') no Confirm before connecting to Internet? (matters only if you use a modem) no Disconnect when done? (matters only if you use a modem) no Connect to network about every 0.0 days On multiprocessors, use at most 1 processors Write to disk at most every 60 seconds Limits on disk space and memory used by BOINC: Use no more than 0.5 GB Leave at least 0.003 GB Use no more than 95% of total space Use no more than 99% of total virtual memory Limits on network bandwidth used by BOINC: Maximum download rate: no limit Maximum upload rate: no limit ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
Send message Joined: 2 Sep 04 Posts: 321 Credit: 10,607 RAC: 0 |
Do work while computer is running on batteries? (matters only for portable computers) no Do work while computer is in use? no Do work only after computer is idle for 3 minutes Do work only between the hours of (no restriction) Leave applications in memory while preempted? (suspended applications will consume swap space if 'yes') no Confirm before connecting to Internet? (matters only if you use a modem) no Disconnect when done? (matters only if you use a modem) no Connect to network about every 3 days On multiprocessors, use at most 1 processors Write to disk at most every 50 seconds Limits on disk space and memory used by BOINC: Use no more than 100 GB Leave at least 0.1 GB Use no more than 85% of total space Use no more than 80% of total virtual memory feel free to visit Guidos Boinc Forum Better to know now if there is a problem than to have an angry mob of users beating at your door.Investigations continue.;-)))Common Criteria (ISO 15408) |
![]() ![]() Send message Joined: 1 Sep 04 Posts: 157 Credit: 82,604 RAC: 0 ![]() |
These are mine: Do work while computer is running on batteries? (matters only for portable computers) no Do work while computer is in use? yes Do work only between the hours of (no restriction) Leave applications in memory while preempted? (suspended applications will consume swap space if 'yes') yes Confirm before connecting to Internet? (matters only if you use a modem) no Disconnect when done? (matters only if you use a modem) no Connect to network about every 1 days On multiprocessors, use at most 2 processors Write to disk at most every 60 seconds Limits on disk space and memory used by BOINC: Use no more than 0.8 GB Leave at least 13 GB Use no more than 8% of total space Use no more than 75% of total virtual memory Limits on network bandwidth used by BOINC: Maximum download rate: no limit Maximum upload rate: no limit Win XP Pro SP1, P4 2.4@2.88GHz Greetings from Belgium |
![]() Send message Joined: 18 Sep 04 Posts: 3 Credit: 13,452 RAC: 0 |
Batteries: Yes, but irrelevant for my machine since it doesn't use batteries Work while computer in use: Yes Work between hours: no restriction Leave apps in memory while preempted: Yes Confirm before connecting: No, but irrelevant because on cable Disconnect when done: No, but still irrlevant Connect to network about every: 2 days Write to disk at most: 60s Disk and memory limits: Use no more than 100GB Leave at least .1 GB Use no more than 50% of total space Use no more than 75% of total virtual memory Bandwidth: No limits Computer that had problems is an AMD Athlon 2400+ (2800+ overclocked) with 512 MB DDR memory (single stick) with a SOYO KT400+ Dragon running WinXP SP2 and BOINC 4.09. |
![]() Send message Joined: 22 Sep 04 Posts: 42 Credit: 2,956 RAC: 0 |
I've had one computation error, when the graphics display crashed LHC4.45, and a number of zero CPU time WUs. My computer: AMD Athlon XP3100+ (Not overclocked), Win98SE (with all the latest Security patches), ~480 MB available RAM. My BOINC settings are set mostly at the SETI@home site: Do work while computer is running on batteries? (matters only for portable computers) yes (Desktop, dosen't run on batteries) Do work while computer is in use? yes Do work only between the hours of (no restriction) Leave applications in memory while preempted? (suspended applications will consume swap space if 'yes') yes (at first, it was set to 'no') Confirm before connecting to Internet? (matters only if you use a modem) no (Connected over 10Mbps LAN) Disconnect when done? (matters only if you use a modem) yes (Connected over 10Mbps LAN) Connect to network about every 5.0 days (Set at SET@home) On multiprocessors, use at most 2 processors (Only 1 processor) Write to disk at most every 60 seconds Limits on disk space and memory used by BOINC: Use no more than 3 GB Leave at least 0.75 GB Use no more than 50% of total space Use no more than 75% of total virtual memory Limits on network bandwidth used by BOINC: Maximum download rate: no limit Maximum upload rate: no limit One other point, I use AMS Fast Defrag 2 Professional (15-Day Demo at http://www.amsn.ro/ ) in an attempt to keep available RAM together. I don't know if it really helps memory leaks, but I can go from 95% "Used Hard Mem" to 21% "Used Hard Mem" when I run it manually. One additional thought... I don't know how the CPU time is kept track of, or displayed. If the CPU time is truncated at two decimal places, or if the CPU time is rounded to two decimal places, and if the actual CPU time is 0.0049 seconds or less, it will show zero CPU Time in a display using only two decimal digits. Is the CPU time kept in the result in more than hundredths of a second? Perhaps another decimal place or two in the display would solve the zero CPU time display problem. Stephen |
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> One additional thought... I don't know how the CPU time is kept track of, or > displayed. If the CPU time is truncated at two decimal places, or if the CPU > time is rounded to two decimal places, and if the actual CPU time is 0.0049 > seconds or less, it will show zero CPU Time in a display using only two > decimal digits. Is the CPU time kept in the result in more than hundredths of > a second? Perhaps another decimal place or two in the display would solve the > zero CPU time display problem. > > Stephen > > the result cpu time seems to be a floating point number with six digits after the decimal point. - http://lhcathome.cern.ch/ v64lhc1000pro44s4_6547.07_1_sixvf_67456_2 446 0 2 2051.218750 0.772240 2051.875000 ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
![]() Send message Joined: 22 Sep 04 Posts: 42 Credit: 2,956 RAC: 0 |
I suspected it was a floating point number, so it could still be a rounding "problem" in the page display. If the duration of the computation was sufficiently short, yet greater than zero, it would display as 0.00 in the displays, which only show two decimal places (hundreths of a second). > > the result cpu time seems to be a floating point number with six digits after > the decimal point. > I'm sorry I took this off topic. Cheers, Stephen |
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> I'm sorry I took this off topic. At least ya didn't turn the thread in to a repub versus democrat flame war. ;) I thought I'd start this thread to see if my settings (ie.. connect every 0.0 days) was the cause of my problems. ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
![]() Send message Joined: 22 Sep 04 Posts: 42 Credit: 2,956 RAC: 0 |
> > At least ya didn't turn the thread in to a repub versus democrat flame war. > ;) > And why would I have done that? We all know the Republicans are CORRECT, and the democrats will continue to be LEFT behind. P-] (Hey, _YOU_ started it! (GRIN)) Why do you have the "Connect to network about every" value set to 0.0? What was the thinking behind that setting? Basically, that setting sets the minimum amount of work to be downloaded based on estimated execution time, while the maximum amount is twice that amount. Cheers, Stephen (You will notice that, unlike some, I have refrained from adding a Bush banner here.) |
![]() Send message Joined: 2 Sep 04 Posts: 126 Credit: 49,653 RAC: 0 |
|
![]() Send message Joined: 22 Sep 04 Posts: 42 Credit: 2,956 RAC: 0 |
Michael, I originally set my "Use no more than" to 10 GB, and "Leave at least" to 1.00 GB. Now, I'm set up with "Use no more than" set at 3 GB, "Leave at least" at 0.75 GB. The BOINC display shows only a small fraction of the available allocation as actually used, and I suspect the 3GB setting may be excessive. I will re-evaluate my disk space settings after Einstein@home starts up. Looking at my numbers, BOINC shows the following: BOINC Software: 12.40MB BOINC Free Space: 2.95GB SET@home: 15.34MB LHC@home: 19.76MB Together, BOINC, SETI, and LHC use around 50MB. I don't process Climate Prediction, but even their large WU's shouldn't take up the additional 2.95GB of disk space. If Alex is short on disk space, I suspect his setting of 500MB/3MB would probably do, unless he also does Climate Prediction. Personally, I would probably double his settings, but I have lots of available HD space. I have no idea of the size of Climate Prediction WU's, other than they are "large." I still question his "Connect to network about every" setting of 0.0 days. My "Connect to network about every" is set to 5.0 days, which implies 5 to 10 days of work for the projects downloaded to my computer. That's long enough to cover any reasonable server outages, yet doesn't get too much work for the project deadlines. I have seen recommendations that 5 days is a good value, while 7 days is, perhaps, too much because of the SETI deadlines. I'm not sure how "0.0" would actually affect the downloading of WU's. I think Alex would get very few WU's, the implication being 0.0 to 0.0 (0.0 x 2 = 0.0) days of work. I just got new work last night (25 Sept '04) for both SETI and LHC, and I haven't yet turned in many results from that download, so my numbers should be close to the "high water mark" for disk space used. As you can see, a "Use no more than" of 3GB leaves lots of available space for SETI and LHC. Currently, both my new SETI and LHC deadlines are about 14 days (09 Oct '04). > These settings could be causing your issues. You'll have very little scratch > space left after downloading applications, support files and result files, > especially if you're participating in more than one project. Try a more > generous allocation of disk space, 10 GB perhaps. > I don't know how much scratch disk space BOINC, SETI, or LHC use. Is it all swapfile? Or are scratch files written, other than the checkpoint files? Cheers, Stephen |
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> > Use no more than 0.5 GB > > Leave at least 0.003 GB > > Alex, > > These settings could be causing your issues. .. I doubt it. IN fact. This app can run off of one of those mini flash drives. It can run under WINE on linux, and I can move it from one computer to the next to the next. Pentium1 to Pentium 2 to AMDk6 to Celeron too. Haven't tried it on my SGI Indy. ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
![]() Send message Joined: 2 Sep 04 Posts: 126 Credit: 49,653 RAC: 0 |
|
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> > I doubt it. > > > > IN fact. This app can run off of one of those mini flash drives. > > > Alex, > > SCRATCH is the key word. It's not about the size of the application. > > [url=http://lhcathome.cern.ch/team_display.php?teamid=45] ![]() ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
![]() ![]() Send message Joined: 2 Sep 04 Posts: 209 Credit: 1,482,496 RAC: 0 |
|
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> > > I'm sorry I took this off topic. > > > > I thought I'd start this thread to see if my settings (ie.. connect every > 0.0 > > days) was the cause of my problems. > > > Sorry, I'm just jumping in here just now. > > What was the original problem ? You message doesn't say. I was wondering why only certain people would notice the 'Boinc "borking" on multiple projects' bug. http://lhcathome.cern.ch/forum_thread.php?id=386 ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
![]() Send message Joined: 22 Sep 04 Posts: 42 Credit: 2,956 RAC: 0 |
> > IN fact. This app can run off of one of those mini flash drives. > > It can run under WINE on linux, and I can move it from one computer to the > next to the next. > Pentium1 to Pentium 2 to AMDk6 to Celeron too. > > Haven't tried it on my SGI Indy. > ... > > Running on knoppix with boinc in a USB flash drive. > Result validated even after being on multiple computers and chipsets. > Alex, Are you moving the same WU from machine to machine? That ought to produce some really bizzar results if you are. I doubt that is even "legal" as far as the project is concerned. Stephen |
![]() Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
> > Are you moving the same WU from machine to machine? That ought to produce some > really bizzar results if you are. I doubt that is even "legal" as far as the > project is concerned. > > Stephen I tried the knoppix thing as a test to see if a work unit would validate.. seems like it did. When they did the 'two day deadline' thing.. I moved a few WU's from my slower machines to my fast machine, results validated ok. Update.. VALIDATED! http://lhcathome.cern.ch/workunit.php?wuid=103560 and for those who want to know how big my flashdrive is.. http://members.shaw.ca/joevoxxer/knoppixboinc.jpg ______________________________________________________________ Did your tech wear a static strap? No? Well, there ya go! :p |
©2025 CERN