Message boards :
Number crunching :
BOINC 7 - problems, solutions and tips
Message board moderation
Author | Message |
---|---|
Cliff Harding Send message Joined: 18 Aug 99 Posts: 1432 Credit: 110,967,840 RAC: 67 |
On 23 January, I successfully upgraded from 6.12.34 -> 7.0.8. Since that time the scheduler has not requested any new tasks. On 1/24 it automatically reported 204 tasks, 1/25 it automatically reported 211 tasks. Today I manually requested an update and reported 150 completed tasks, expecting to receive a number of tasks, but the BM requested none. I am down to approx. 4 days of work left on this machine. Prior to the upgrade, the system normally kept at least 10 days in the queue. There are three projects on the machine (SETI, Milkyway & Einstein), with only SETI allowed to get tasks. Is there something on the server side that is severly restricting any machine from getting a full allotment of work when it applies to 7.xx.xx versions? The machine, A-SYS, is running W7/64-bit, 4Gb ram, Lunatics v.039, (1) EVGA GTX460 SE, (1) EVGA GTS250. I don't buy computers, I build them!! |
LadyL Send message Joined: 14 Sep 11 Posts: 1679 Credit: 5,230,097 RAC: 0 |
Please post here if you are running BOINC 7 (Alpha!) and are having issues. The scheduling in BOINC 7 has been revamped, you may be seeing that. Also if you encounter bugs they need to be fed back to the devs. IF you are running BOINC alpha clients you should be reading the mailing list! |
LadyL Send message Joined: 14 Sep 11 Posts: 1679 Credit: 5,230,097 RAC: 0 |
quoting from a different thread Workfetch will be initiated when the calculated remaining work drops below the 'connect every' setting and will then ask for as much work as specified in the 'aditional work' setting. It will ask from the project for which the debt entry is the smallest (calculated as work done recently compared to resource share). If no work is available from that project it will ask from the next larger and so on until it gets work. @Cliff If you can only fetch from SETI, you probably have more cache left (over all projects) than whatever your 'connect every' setting is. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
Hi, I think scheduler is great except when running too many projects... It seems like some projects gets behind when i ran more then 4 projects on 1 ATI GPU. The fifth project never started....Maybe I should have waited more then 1 week. I now run some SETI Beta on my ATI cards 5850 + 4850 and i haven't set the flops in app_info.xml so the tasks has very long runtimes when they start so I only have 2 tasks i que. But sceduling works very nice and when 1 task only has a few hours left it gets a new task. Great!!! A not nessesary wish would be to let more projects to run at the same time(alternating and running) at the same time. When it works now it "feels" very good for me as a user. No more suspend and NNT in boincmanager. TRuEQ & TuVaLu |
Piotr Kunkel Send message Joined: 7 Apr 00 Posts: 18 Credit: 19,385,083 RAC: 0 |
Cliff Harding - try 7.02, is seems to be good. Is's, for me, the only 7.xx whitch requesting and downloading new tasks. " Then they came for me and there was no one left to speak out for me." Martin Niemöller |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
Is there any change that the message button(event log) can be put back next to the disk option in BoincManager? TRuEQ & TuVaLu |
iwazaru Send message Joined: 31 Oct 99 Posts: 173 Credit: 509,430 RAC: 0 |
Any idea why this WU resulted in an error? Apologies in advance if not a Boinc 7(.08) problem... http://setiathome.berkeley.edu/workunit.php?wuid=916898344 |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Any idea why this WU resulted in an error? Apologies in advance if not a Boinc 7(.08) problem... Well, the task details report <message> upload failure: <file_xfer_error> <file_name>18oc11aa.365.12101.9.10.62.vlar_0_0</file_name> <error_code>-131</error_code> </file_xfer_error> </message> In other words, the task finished normally, but there was a communications glitch afterwards when it tried to send the outcome back to Berkeley. It's possible that is related to BOINC v7.0.8, but unlikely - the preceding task uploaded properly, and that was with v7.0.8, too. How long have you been running that particular (experimental) version - in other words, how many tasks have you completed since the upgrade? If it's been OK for a while, and this is the first error, I'd be inclined to put it down to a random gremlin, and move on. However, v7.0.8 is a bit old to still be testing - as the thread title now says, testing is now up to v7.0.12. That seems to be running well here, so you might feel inclined to upgrade and help us find a whole new set of bugs ;-) |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
<error_code>-131</error_code> Error -131 means that the size of one of the output files was bigger than the maximum set by the project for upload. BOINC will not try to upload this file. Perhaps something to do with those mega MBs? |
iwazaru Send message Joined: 31 Oct 99 Posts: 173 Credit: 509,430 RAC: 0 |
...testing is now up to v7.0.12. That seems to be running well here, so you might feel inclined to upgrade and help us find a whole new set of bugs ;-) Sure! Where does 7.0.12 live?:-) My Seti knowledge only takes me this far: http://boinc.berkeley.edu/download_all.php PS I've retired this 7 year old 73W Celeron for energy/efficiency reasons, but it never gave me any trouble, so it should be OK (keeping an eye on it, nevertheless). Just wanted to see what Boinc 7 looked like... Edit: OK, I seem to have found the source of my confusion. The second post in the thread looks like it should have been the first. I've never seen that in the forums before... So what/where is Boinc Alpha? |
Wembley Send message Joined: 16 Sep 09 Posts: 429 Credit: 1,844,293 RAC: 0 |
Try here: http://boinc.berkeley.edu/dl/?C=M;O=D |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
<error_code>-131</error_code> It's something to watch out for when we crunch them, but I doubt it. I've got several mega MBs now, but every single one of them is called 21jn11ac.5207... - the tasks produced by other splitter instances working on the same 21jn11ac 'tape' are normal size. iwazaru's file is called 18oc11aa.365..., and has been successfully completed by the other user, so I think it must be a different issue. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
I abandoned the getting my fifth project Albert to get it's schedular prio the same as the other 4 projects that worked together as they should. Now I am running 3 projects SETI Beta , SETI and Albert. The problem is Albert is showing the same difference in scheduling prio as it did when i ran more projects. Could this have to do with Albert in any way?? TRuEQ & TuVaLu |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
Ok, now Albert is alternating with Milkyway, SETI and SETI Beta. Seti is set to use only ap tasks which is not available at the moment. Seti Beta is set to NNT and completed it's last three tasks and then had 0 tasks in cue. Milkyway is running and then Albert prio for some reason got closer to the others. Maybe it just needed to get going with it's big difference prio to start functioning as it should. SETI is still the project that should be running according to it's prio but Albert and Milkyway are in fact the once that are alternating and running as they should since SETI ap tasks are not available. It was like prio calculation for Albert got stuck with a difference from the other projects. It did go up and down but not enough to make Albert run. TRuEQ & TuVaLu |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
A tip. Poem has like 0 or 1 task avalable from time to time for their GPU app. My BM(7.0.11) doesn't seem to "click on update" when the "communication deffered" time has come to 0 So, I don't get any tasks from time to time. When I click on the update when it's not deffered a couple of times I usually get 1 or 2 tasks. Couldn't the BM do this?? note: POEM has a very short "communication deffered". TRuEQ & TuVaLu |
MarkJ Send message Joined: 17 Feb 08 Posts: 1139 Credit: 80,854,192 RAC: 5 |
There are issues around gpu work fetch with the current alpha versions (including 7.0.12). The BOINC alpha mailing list is where we discuss these things. |
TRuEQ & TuVaLu Send message Joined: 4 Oct 99 Posts: 505 Credit: 69,523,653 RAC: 10 |
There are issues around gpu work fetch with the current alpha versions (including 7.0.12). The BOINC alpha mailing list is where we discuss these things. Ok, then I know you guys are working on it. TRuEQ & TuVaLu |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.