getting WU not applied for
24 Mar 2012 16:33:44 UTC
Topic 196252
(moderation:
I've recently reset E@H due to too many tasks due to time out before completion.
Didnt want to waste cpu time and wingmans time by waiting them out.
I went through my prefs and de selected GWs6lineVeto WU's, but guess what I now have a good few anyway.
However I cannot get any GPU apps..
Server keeps saying check your prefs, the only WU not selected dont relate AFAIK to GPU tasks.
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
getting WU not applied for
)
Have a look at the server log:
2012-03-24 16:29:46.9803 [PID=28924] Request: [USER#xxxxx] [HOST#4695229] [IP xxx.xxx.xxx.72] client 6.12.34
2012-03-24 16:29:46.9894 [PID=28924] [send] effective_ncpus 6 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
2012-03-24 16:29:46.9894 [PID=28924] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
2012-03-24 16:29:46.9894 [PID=28924] [send] Not using matchmaker scheduling; Not using EDF sim
2012-03-24 16:29:46.9894 [PID=28924] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2012-03-24 16:29:46.9894 [PID=28924] [send] CUDA: req 65028.44 sec, 2.00 instances; est delay 0.00
2012-03-24 16:29:46.9895 [PID=28924] [send] work_req_seconds: 0.00 secs
2012-03-24 16:29:46.9895 [PID=28924] [send] available disk 90.36 GB, work_buf_min 21600
2012-03-24 16:29:46.9895 [PID=28924] [send] active_frac 0.999920 on_frac 0.973642 DCF 1.000000
2012-03-24 16:29:47.0029 [PID=28924] [version] Checking plan class 'BRP4cuda32'
2012-03-24 16:29:47.0033 [PID=28924] [version] reading plan classes from file '../plan_class_spec.xml'
2012-03-24 16:29:47.0034 [PID=28924] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2012-03-24 16:29:47.0034 [PID=28924] [version] driver version required max: -29053, supplied: 29573
2012-03-24 16:29:47.0034 [PID=28924] [version] Checking plan class 'BRP4SSE'
2012-03-24 16:29:47.0034 [PID=28924] [version] parsed project prefs setting 'also_run_cpu' : true : 0.000000
2012-03-24 16:29:47.0034 [PID=28924] [version] Don't need CPU jobs, skipping version 122 for einsteinbinary_BRP4 (BRP4SSE)
2012-03-24 16:29:47.0034 [PID=28924] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
2012-03-24 16:29:47.0034 [PID=28924] [CRITICAL] [HOST#4695229] can't resend [RESULT#279374250]: no app version for einsteinbinary_BRP4
2012-03-24 16:29:47.0041 [PID=28924] [CRITICAL] [HOST#4695229] can't resend [RESULT#279374255]: no app version for einsteinbinary_BRP4
Then read this message from the developers and the one immediately before it.
You need to revert to an earlier NVidia driver on host 4695229.
Hi Richard, Well
)
Hi Richard,
Well spotted that man:-) I'll have a look at my store of nvidia gforce drivers and see if I still have 285.nn
Now that explains why no cuda WU but what explains the not selected WU's?
Regards,
Cliff,
Been there, Done that, Still no damm T Shirt.
RE: Hi Richard,
)
Because you didn't need any. CPU: req 0.00 sec, 0.00 instances; est delay 0.00
You can always download the older drivers from http://www.nvidia.co.uk/Download/Find.aspx?lang=en-uk
RE: Now that explains why
)
What I've been told is that the Gravitational Wave Search is what pays E@H's bills (being the work that their grants are for), so even if you de-select the GW or LineVeto apps, they'll still continue to push them to clients.
In my personal case, I found it a bit obnoxious, since on one of my machines it's a complete and total waste of bandwidth, given that they all error out pretty much instantly.