Message boards :
Number crunching :
Help please...
Message board moderation
Author | Message |
---|---|
Squyrrel Send message Joined: 8 Feb 00 Posts: 14 Credit: 24,882,903 RAC: 0 |
I ventured into the world of optimised apps and it failed. While I was stupid enough not to take a 'proper' backup of the BOINC folder structure, I did manage to recover most of it using an undelete prog. I've tried a numerous ways to inject the old work and config but to no avail. Is there a way to manually add work files into the client? If so could I get some guidance on how to please? Thanks. Win 7/64 NV275gtx E8400 BOINC 6.10.17 |
DJStarfox Send message Joined: 23 May 01 Posts: 1066 Credit: 1,226,053 RAC: 2 |
Is there a way to manually add work files into the client? No. Those workunits will take care of themselves on the server-side. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
It's almost impossible. For each task, you would have to create two blocks of data in client_state.xml, like these: <workunit> <name>01ap07aa.4377.481.12.10.139</name> <app_name>setiathome_enhanced</app_name> <version_num>603</version_num> <rsc_fpops_est>166495399315177.000000</rsc_fpops_est> <rsc_fpops_bound>1664953993151770.000000</rsc_fpops_bound> <rsc_memory_bound>33554432.000000</rsc_memory_bound> <rsc_disk_bound>33554432.000000</rsc_disk_bound> <file_ref> <file_name>01ap07aa.4377.481.12.10.139</file_name> <open_name>work_unit.sah</open_name> </file_ref> </workunit> <result> <name>01ap07aa.4377.481.12.10.139_1</name> <final_cpu_time>0.000000</final_cpu_time> <exit_status>0</exit_status> <state>2</state> <platform>windows_intelx86</platform> <version_num>603</version_num> <wu_name>01ap07aa.4377.481.12.10.139</wu_name> <report_deadline>1261419506.000000</report_deadline> <file_ref> <file_name>01ap07aa.4377.481.12.10.139_1_0</file_name> <open_name>result.sah</open_name> </file_ref> </result> (and those are CPU tasks. CUDA is even more complicated). You've probably undeleted the actual task data files, so you could plug those names in. But you would need to refer to the website to find out which replication you got - those _0 or _1 at the end. And to do the job properly, you'd need to translate the deadlines off the website into Unix standard form. In your post at Lunatics, you said it was your Cuda box which had suffered this misfortune. Your CUDA host, 5126909, has 744 tasks in progress, dating back to 1 Oct 2009 11:10:24 UTC. (How? Why?) You are going to have very tired fingers and eyes by the time you've finished stitching that lot together, and the chance of doing it without a single error are virtually nil. Oh, and did I say that BOINC will have to remain idle throughout the process? No, best to cut your losses, take your medicine and call it a lesson learned. If you have any tasks running at the moment, let them finish, but don't allow any new work to download. Then detach the host from the project, and re-attach it - that will release the tasks which have now become 'ghosts' back for the community to finish off, while we're waiting for new work. Finally, with 'No new tasks' still set, complete the optimisation job you set out to do in the first place (making sure you tick the CUDA box this time!), and just download a few tasks at first to test that it's working before you unleash the floodgates on the massive cache again. |
Squyrrel Send message Joined: 8 Feb 00 Posts: 14 Credit: 24,882,903 RAC: 0 |
Thanks Richard All my problems solved now. I was a bit miffed at the loss really only because there was no work coming through from Seti. The results are back in the farm for redistribution already anyway. I also found a suitable answer to the root cause; which is due to the way the app handles bitness, ie 32/64. I stumbled on the answer over at lunatics courtesy of Josef... http://lunatics.kwsn.net/windows/help-i-guess.0.html It may be useful for that gem to be easier to find or at least the next installer version make the change... Please mark solved. Thanks. |
©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.