Nvidia GT630 always fails after 3 seconds on (Arecibo, GPU) v1.39 (BRP4G-cuda32-nv301)
1 Sep 2013 8:55:05 UTC
Topic 197176
(moderation:
GT630 is basically a GT 440 with 2 GB of ddr3 memory. My system with a GT 440 succeeds but the GT 630 fails with the following error:
Quote:
7.2.11
The system cannot find the file specified.
(0x2) - exit code 2 (0x2)Activated exception handling...
[02:49:11][4004][INFO ] Starting data processing...
[02:49:11][4004][INFO ] CUDA global memory status (initial GPU state, including context):
------> Used in total: 79 MB (1970 MB free / 2049 MB total) -> Used by this application (assuming a single GPU task): 0 MB
[02:49:11][4004][INFO ] Using CUDA device #0 "GeForce GT 630" (96 CUDA cores / 311.04 GFLOPS)
[02:49:11][4004][INFO ] Version of installed CUDA driver: 5050
[02:49:11][4004][INFO ] Version of CUDA driver API used: 3020
[02:49:11][4004][ERROR] Header checkpoint file status.cpt contains inconsistent information about number of templates done (236917 > 6662).
[02:49:11][4004][ERROR] Demodulation failed (error: 2)!
02:49:11 (4004): called boinc_finish
On the GT440, which succeeds with the task, this section says:
Quote:
7.2.5
Activated exception handling...
[04:05:30][3348][INFO ] Starting data processing...
[04:05:30][3348][INFO ] CUDA global memory status (initial GPU state, including context):
------> Used in total: 129 MB (896 MB free / 1025 MB total) -> Used by this application (assuming a single GPU task): 0 MB
[04:05:30][3348][INFO ] Using CUDA device #0 "GeForce GT 440" (96 CUDA cores / 315.84 GFLOPS)
[04:05:30][3348][INFO ] Version of installed CUDA driver: 5050
[04:05:30][3348][INFO ] Version of CUDA driver API used: 3020
[04:05:30][3348][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory).
------> Starting from scratch...
Any idea what's causing the problem on the GT630? It works fine on other projects (Seti, milkyway, poem, Primegrid). Could it be that the GT630 machine is running Boinc 7.2.11 while the GT440 machine is running Boinc 7.2.5?
Thanks,
David Ball
Nvidia GT630 always fails after 3 seconds on (Arecibo, GPU) v1.3
)
OK, after at least 14 workunits failing with that error, the next one I got seems to be working. I did notice that in one of the 14 failures it ran for 21 seconds and the error was slightly different. In that one, the status.cpt had the wrong task name in it.
The task name was PA0020_00621_22_0
-- David Ball
I'm still getting a mix of
)
I'm still getting a mix of errors and an occasional success on that machine. All other GPU projects are successful. Any idea where the errors are in the status.cpt ?
Did you try a file system
)
Did you try a file system check already? Perhaps the internal pointers in the project or slots directory got corrupted.
Gruß
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
