Discussed before, But ANWAY, HUGE problem!
22 Mar 2012 2:06:16 UTC
Topic 196247
(moderation:
Hi!
I'm aware with the NVIDIA CUDA problems with the newest drivers, but with other GPU projects I have NO problems at all!
I use WIN7 64 bit, BOINC 7.0.22, GeForce GTX 550 Ti, driver 296.10 (official, not BETA) have ABSOLUTELY no problems with GPUGRID, Milkway and SETI GPU WUs, since I configured the system NOT to go to sleep EVER!
Yet, Einstein refuses me (the previous driver, I think 295.. something did just fine?!) to send ANY GPU WUs, with explanation to look on a particular link, which states:
2012-03-22 01:21:23.2505 [PID=5081] Request: [USER#xxxxx] [HOST#4281670] [IP xxx.xxx.xxx.88] client 7.0.22
2012-03-22 01:21:23.2581 [PID=5081 ] [send] effective_ncpus 2 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
2012-03-22 01:21:23.2581 [PID=5081 ] [send] effective_ngpus 1 max_jobs_on_host_gpu 999999
2012-03-22 01:21:23.2581 [PID=5081 ] [send] Not using matchmaker scheduling; Not using EDF sim
2012-03-22 01:21:23.2581 [PID=5081 ] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2012-03-22 01:21:23.2581 [PID=5081 ] [send] CUDA: req 864180.00 sec, 1.00 instances; est delay 0.00
2012-03-22 01:21:23.2581 [PID=5081 ] [send] work_req_seconds: 0.00 secs
2012-03-22 01:21:23.2582 [PID=5081 ] [send] available disk 43.62 GB, work_buf_min 0
2012-03-22 01:21:23.2582 [PID=5081 ] [send] active_frac 0.993184 on_frac 0.992323 DCF 0.900305
2012-03-22 01:21:23.2589 [PID=5081 ] [send] [HOST#4281670] is reliable
2012-03-22 01:21:23.2589 [PID=5081 ] [send] set_trust: random choice for error rate 0.011447: no
2012-03-22 01:21:23.3953 [PID=5081 ] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket ()
2012-03-22 01:21:23.3953 [PID=5081 ] [version] Checking plan class 'SSE2'
2012-03-22 01:21:23.3957 [PID=5081 ] [version] reading plan classes from file '../plan_class_spec.xml'
2012-03-22 01:21:23.3958 [PID=5081 ] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket (SSE2)
2012-03-22 01:21:23.3958 [PID=5081 ] [version] Checking plan class 'SSE'
2012-03-22 01:21:23.3958 [PID=5081 ] [version] Don't need CPU jobs, skipping version 102 for einstein_S6Bucket (SSE)
2012-03-22 01:21:23.3958 [PID=5081 ] [version] no app version available: APP#16 (einstein_S6Bucket) PLATFORM#2 (windows_intelx86) min_version 0
2012-03-22 01:21:23.4082 [PID=5081 ] [version] Checking plan class 'BRP4cuda32'
2012-03-22 01:21:23.4082 [PID=5081 ] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2012-03-22 01:21:23.4083 [PID=5081 ] [version] driver version required max: -29053, supplied: 29610
2012-03-22 01:21:23.4083 [PID=5081 ] [version] Checking plan class 'BRP4SSE'
2012-03-22 01:21:23.4083 [PID=5081 ] [version] parsed project prefs setting 'also_run_cpu' : true : 0.000000
2012-03-22 01:21:23.4083 [PID=5081 ] [version] Don't need CPU jobs, skipping version 122 for einsteinbinary_BRP4 (BRP4SSE)
2012-03-22 01:21:23.4083 [PID=5081 ] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
2012-03-22 01:21:23.4083 [PID=5081 ] [version] Don't need CPU jobs, skipping version 23 for hsgamma_FGRP1 ()
2012-03-22 01:21:23.4083 [PID=5081 ] [version] no app version available: APP#17 (hsgamma_FGRP1) PLATFORM#2 (windows_intelx86) min_version 0
2012-03-22 01:21:23.4117 [PID=5081 ] [send] [HOST#4281670] is looking for work from a non-preferred application
2012-03-22 01:21:23.4165 [PID=5081 ] [debug] [HOST#4281670] MSG(high) No work sent
2012-03-22 01:21:23.4166 [PID=5081 ] [debug] [HOST#4281670] MSG(high) see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/4281/4281670
2012-03-22 01:21:23.4166 [PID=5081 ] Sending reply to [HOST#4281670]: 0 results, delay req 60.00
2012-03-22 01:21:23.4169 [PID=5081 ] Scheduler ran 0.172 seconds
If the driver is the problem, why did the previous driver worked just fine with Einstein GPU tasks (and it has been said, that it is also problematic)?
If the driver is the culprit, is there ANY other way way as substitute it with the really OLD one?
Sorry, if this has been already answered, but I had not the time to search for answers!
Thank you.
Denis
“A little knowledge is a dangerous thing. So is a lot.” - Albert EINSTEIN
Discussed before, But ANWAY, HUGE problem!
)
Your required action is shown by this line you posted.
I suggest you take the required action. If you wish not to do so, I suggest that you could disable GPU use on this project.
As to searching for answers, one would not have to look very far--something like half the recent thread starts are duplications of this topic.
Oliver Bock wrote:FYI, we
)
(Post 116397)
And
(Post 116415)
And
Thread (9354)
Hi! Thank you all for the
)
Hi!
Thank you all for the answers I already expected (assumed). But better to ask and get the answers from people who are actually part of this and trained to provide answers, which are based on facts not on assumptions.
Best wishes.
Denis
“A little knowledge is a dangerous thing. So is a lot.” - Albert EINSTEIN