SOS! Einstein project.
I have a troubles with GPU-using tasks during last 12 days.
All GPU-using tasks is totally failed with error code: "Error while computing"
Binary Radio Pulsar Search v1.05 (BRP3SSE) -- All tasks OK (4 tasks)
Global Correlations S5 HF search #1 v1.07 (SSE2) -- All tasks OK (1 task)
Binary Radio Pulsar Search v1.08 (BRP3cuda32fullCPU) -- ALL tasks totally failed: Error while computing (63 tasks)
My PC:
CPU type GenuineIntel
Intel(R) Core(TM)2 Duo CPU E7500 @ 2.93GHz [Family 6 Model 23 Stepping 10]
Number of processors 2
Coprocessors NVIDIA GeForce 210 (511MB)
Operating System Mandriva Linux release 2010.2 (Official) for i586
2.6.33.7-desktop586-2mnb
BOINC client version 6.10.58
Memory 2023.48 MB
Card: NVIDIA GeForce 6100 to GeForce 360
Driver: nvidia 260.19.44
I have no idea what is happening. Pls help resolve this puzzle if it possible!
I suspect this troubles is result of update of system which I have ~3 weeks ago.
Mandriva 2010.1 -> 2010.2
A lot of components of system was changed during this update.
Maybe something was wrong for BOINC...
Is it possible to disable GPU using fo BOINC? How can I disable GPU using?
I have two kind of major errors:
Exit status 252 (0xfc)
6.10.58
process exited with code 252 (0xfc, -4)
[19:13:43][2677][INFO ] Starting data processing...
[19:13:43][2677][ERROR] Couldn't initialize CUDA driver API (error: 100)!
[19:13:43][2677][ERROR] Demodulation failed (error: 1020)!
19:13:43 (2677): called boinc_finish
]]>
and
Exit status -226 (0xffffffffffffff1e)
[19:06:51][19017][ERROR] Demodulation failed (error: 1006)!
[19:06:51][19017][WARN ] CUDA memory allocation problem encountered!
------> Returning control to BOINC, delaying restart for at least five minutes...
------> If this problem persists you should consider aborting this task.
[19:06:51][19032][INFO ] Starting data processing...
[19:06:51][19032][INFO ] CUDA global memory status (initial GPU state, including context):
------> Used in total: 296 MB (216 MB free / 512 MB total) -> Used by this application (assuming a single GPU task): 0 MB
[19:06:51][19032][INFO ] Using CUDA device #0 "GeForce 210" (16 CUDA cores / 67.30 GFLOPS)
[19:06:51][19032][INFO ] Version of installed CUDA driver: 3020
[19:06:51][19032][INFO ] Version of CUDA driver API used: 3020
[19:06:51][19032][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory).
------> Starting from scratch...
[19:06:51][19032][INFO ] Header contents:
------> Original WAPP file: /BOINC/projects/EinsteinAtHome/temp_working/BRP3/PM0084_04681/PM0084_04681_DM96.00
------> Sample time in microseconds: 500
------> Observation time in seconds: 2099.8515
------> Time stamp (MJD): 51372.223608757209
------> Number of samples/record: 0
------> Center freq in MHz: 1231.5
------> Channel band in MHz: 3
------> Number of channels/record: 96
------> Nifs: 1
------> RA (J2000): 104446.611599
------> DEC (J2000): -543522.559999
------> Galactic l: 0
------> Galactic b: 0
------> Name: G4679519
------> Lagformat: 0
------> Sum: 1
------> Level: 3
------> AZ at start: 0
------> ZA at start: 0
------> AST at start: 0
------> LST at start: 0
------> Project ID: --
------> Observers: --
------> File size (bytes): 0
------> Data size (bytes): 0
------> Number of samples: 4194304
------> Trial dispersion measure: 96 cm^-3 pc
------> Scale factor: 16.9333
[19:06:53][19032][INFO ] Seed for random number generator is -1069745676.
[19:06:55][19032][ERROR] Error allocating power spectrum device memory: 25166336 bytes (error: 2)
[19:06:55][19032][ERROR] Demodulation failed (error: 1006)!
[19:06:55][19032][WARN ] CUDA memory allocation problem encountered!
------> Returning control to BOINC, delaying restart for at least five minutes...
------> If this problem persists you should consider aborting this task.
To disable Einstein@Home GPU tasks for all your PCs, go to http://einstein.phys.uwm.edu/prefs.php?subset=project and clock on "Edit Einstein@Home preferences". On the following dialog, you can disable GPU processing.
It seems like all of my GPU 'cuda' tasks end with computation error. I've adjusted the pref's to not send GPU tasks, but I don't see this error with other projects.
Task 233377580 is an example.
Have you tried the first, simplest thing in the book, when it comes to problems with the videocard: A reboot of the system? That'll clear up many problems one has with videocards and their GPUs. Stuck tasks and such...
It's also possible that your videocard driver got corrupt and needs reinstallation. But try a simple reboot first.
Have you tried the first, simplest thing in the book, when it comes to problems with the videocard: A reboot of the system? That'll clear up many problems one has with videocards and their GPUs. Stuck tasks and such...
It's also possible that your videocard driver got corrupt and needs reinstallation. But try a simple reboot first.
What Jord didn't say, and is probably not needed but here it is anyway, a video card will not reset unless you physically reboot the system. So if the video card got wonky then the ONLY way to reset is to reboot the pc. Now returning you to your regular, and the guy who knows much more than I do, Technical Help guy!
Also notice the difference between a restart and a reboot. Some BOINC projects have had problems that wouldn't go away unless you did a full shutdown, then waited at least 5 minutes before turning the computer on again so it will do a full reboot.
I get a Computation Error about 2 seconds after an S6Bucket job starts. I am running a I7-980 on an ASUS P6T7 WS Supercomputer MB with Tesla C2050's. All other E@H applications run fine. Milkyway, PrimeGrid and others run fine too. I have deselected S6 work from preferences but they keep comming in. Any idease out there? Stargal ( Liz Moore)
RE: Resetting the project
)
Did you reboot (power cycle) that host since 2011-04-07? Perhaps the GPU (memory) got stuck.
There are no erroneous tasks listed for your Q9300. Didn't you report them yet?
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
SOS! Einstein project. I
)
SOS! Einstein project.
I have a troubles with GPU-using tasks during last 12 days.
All GPU-using tasks is totally failed with error code: "Error while computing"
Binary Radio Pulsar Search v1.05 (BRP3SSE) -- All tasks OK (4 tasks)
Global Correlations S5 HF search #1 v1.07 (SSE2) -- All tasks OK (1 task)
Binary Radio Pulsar Search v1.08 (BRP3cuda32fullCPU) -- ALL tasks totally failed: Error while computing (63 tasks)
My PC:
CPU type GenuineIntel
Intel(R) Core(TM)2 Duo CPU E7500 @ 2.93GHz [Family 6 Model 23 Stepping 10]
Number of processors 2
Coprocessors NVIDIA GeForce 210 (511MB)
Operating System Mandriva Linux release 2010.2 (Official) for i586
2.6.33.7-desktop586-2mnb
BOINC client version 6.10.58
Memory 2023.48 MB
Card: NVIDIA GeForce 6100 to GeForce 360
Driver: nvidia 260.19.44
I have no idea what is happening. Pls help resolve this puzzle if it possible!
I suspect this troubles is result of update of system which I have ~3 weeks ago.
Mandriva 2010.1 -> 2010.2
A lot of components of system was changed during this update.
Maybe something was wrong for BOINC...
Is it possible to disable GPU using fo BOINC? How can I disable GPU using?
I have two kind of major errors:
Exit status 252 (0xfc)
6.10.58
process exited with code 252 (0xfc, -4)
[19:13:43][2677][INFO ] Starting data processing...
[19:13:43][2677][ERROR] Couldn't initialize CUDA driver API (error: 100)!
[19:13:43][2677][ERROR] Demodulation failed (error: 1020)!
19:13:43 (2677): called boinc_finish
]]>
and
Exit status -226 (0xffffffffffffff1e)
[19:06:51][19017][ERROR] Demodulation failed (error: 1006)!
[19:06:51][19017][WARN ] CUDA memory allocation problem encountered!
------> Returning control to BOINC, delaying restart for at least five minutes...
------> If this problem persists you should consider aborting this task.
[19:06:51][19032][INFO ] Starting data processing...
[19:06:51][19032][INFO ] CUDA global memory status (initial GPU state, including context):
------> Used in total: 296 MB (216 MB free / 512 MB total) -> Used by this application (assuming a single GPU task): 0 MB
[19:06:51][19032][INFO ] Using CUDA device #0 "GeForce 210" (16 CUDA cores / 67.30 GFLOPS)
[19:06:51][19032][INFO ] Version of installed CUDA driver: 3020
[19:06:51][19032][INFO ] Version of CUDA driver API used: 3020
[19:06:51][19032][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory).
------> Starting from scratch...
[19:06:51][19032][INFO ] Header contents:
------> Original WAPP file: /BOINC/projects/EinsteinAtHome/temp_working/BRP3/PM0084_04681/PM0084_04681_DM96.00
------> Sample time in microseconds: 500
------> Observation time in seconds: 2099.8515
------> Time stamp (MJD): 51372.223608757209
------> Number of samples/record: 0
------> Center freq in MHz: 1231.5
------> Channel band in MHz: 3
------> Number of channels/record: 96
------> Nifs: 1
------> RA (J2000): 104446.611599
------> DEC (J2000): -543522.559999
------> Galactic l: 0
------> Galactic b: 0
------> Name: G4679519
------> Lagformat: 0
------> Sum: 1
------> Level: 3
------> AZ at start: 0
------> ZA at start: 0
------> AST at start: 0
------> LST at start: 0
------> Project ID: --
------> Observers: --
------> File size (bytes): 0
------> Data size (bytes): 0
------> Number of samples: 4194304
------> Trial dispersion measure: 96 cm^-3 pc
------> Scale factor: 16.9333
[19:06:53][19032][INFO ] Seed for random number generator is -1069745676.
[19:06:55][19032][ERROR] Error allocating power spectrum device memory: 25166336 bytes (error: 2)
[19:06:55][19032][ERROR] Demodulation failed (error: 1006)!
[19:06:55][19032][WARN ] CUDA memory allocation problem encountered!
------> Returning control to BOINC, delaying restart for at least five minutes...
------> If this problem persists you should consider aborting this task.
Hi! To disable
)
Hi!
To disable Einstein@Home GPU tasks for all your PCs, go to
http://einstein.phys.uwm.edu/prefs.php?subset=project and clock on "Edit Einstein@Home preferences". On the following dialog, you can disable GPU processing.
HB
Thank you!
)
Thank you!
It seems like all of my GPU
)
It seems like all of my GPU 'cuda' tasks end with computation error. I've adjusted the pref's to not send GPU tasks, but I don't see this error with other projects.
Task 233377580 is an example.
Have you tried the first,
)
Have you tried the first, simplest thing in the book, when it comes to problems with the videocard: A reboot of the system? That'll clear up many problems one has with videocards and their GPUs. Stuck tasks and such...
It's also possible that your videocard driver got corrupt and needs reinstallation. But try a simple reboot first.
RE: Have you tried the
)
What Jord didn't say, and is probably not needed but here it is anyway, a video card will not reset unless you physically reboot the system. So if the video card got wonky then the ONLY way to reset is to reboot the pc. Now returning you to your regular, and the guy who knows much more than I do, Technical Help guy!
Also notice the difference
)
Also notice the difference between a restart and a reboot. Some BOINC projects have had problems that wouldn't go away unless you did a full shutdown, then waited at least 5 minutes before turning the computer on again so it will do a full reboot.
I get a Computation Error
)
I get a Computation Error about 2 seconds after an S6Bucket job starts. I am running a I7-980 on an ASUS P6T7 WS Supercomputer MB with Tesla C2050's. All other E@H applications run fine. Milkyway, PrimeGrid and others run fine too. I have deselected S6 work from preferences but they keep comming in. Any idease out there? Stargal ( Liz Moore)
RE: I have deselected S6
)
You can't deselect them.
Did you already google the exit code -1073741819 (0xc0000005)?
Sounds like an access violation to me. Did you check your AV error logs?
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
