CUDA Errors

Author	Message
Danny Sosebee Send message Joined: 10 Jun 02 Posts: 53 Credit: 1,022,592 RAC: 0	Message 873816 - Posted: 8 Mar 2009, 19:57:04 UTC Last modified: 8 Mar 2009, 19:58:51 UTC While exploring my completed work unit information on the S@H web site I noticed that all of the MB workunits I have processed (using CUDA) have the following error lines in them: Cuda error 'cudaMallocArray( &dev_gauss_dof_lcgf_cache' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcc_gaussfit.cu' in line 417 : feature is not yet implemented. setiathome_CUDA: CUDA runtime ERROR in device memory allocation (Step 2 of 3). Falling back to HOST CPU processing... This particular one is from Task # 1174271269 I am using a NVidia GeForce 8600GT w/512MB, I am not overclocking and am using the stock SETI applications. The drivers were the ones that came with my video card (sorry I do not remember the version number), however I have just now upgraded to Forceware 180.21 (I think that's the version number, it's the latest update on NVidia's website). Am currently running 2 AP tasks and no CUDA, but was wondering if this indicates some kind of problem? Oh yes, OS is Windows XP SP3 (32 bit). ID: 873816 ·

BigWaveSurfer Send message Joined: 29 Nov 01 Posts: 186 Credit: 36,311,381 RAC: 141	Message 873850 - Posted: 8 Mar 2009, 21:50:16 UTC - in response to Message 873816. Actually, the most current drivers listed for your card are 182.08 http://www.nvidia.com/object/winvista_x86_182.08_whql.html Not sure about the other stuff though. ID: 873850 ·

Jord Volunteer tester Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3	Message 873871 - Posted: 8 Mar 2009, 22:40:24 UTC - in response to Message 873816. Cuda error 'cudaMallocArray( &dev_gauss_dof_lcgf_cache' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcc_gaussfit.cu' in line 417 : feature is not yet implemented. setiathome_CUDA: CUDA runtime ERROR in device memory allocation (Step 2 of 3). Falling back to HOST CPU processing... It's telling you there that it has a problem with the memory of your video card. Have you tried to reboot, yet? Or otherwise exited BOINC and restarted it? ID: 873871 ·

Danny Sosebee Send message Joined: 10 Jun 02 Posts: 53 Credit: 1,022,592 RAC: 0	Message 873918 - Posted: 8 Mar 2009, 23:50:28 UTC - in response to Message 873871. Last modified: 9 Mar 2009, 0:50:04 UTC Cuda error 'cudaMallocArray( &dev_gauss_dof_lcgf_cache' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcc_gaussfit.cu' in line 417 : feature is not yet implemented. setiathome_CUDA: CUDA runtime ERROR in device memory allocation (Step 2 of 3). Falling back to HOST CPU processing... It's telling you there that it has a problem with the memory of your video card. Have you tried to reboot, yet? Or otherwise exited BOINC and restarted it? Rebooted several times, but also often run many other programs besides BOINC including an online role playing game (I always suspend BOINC processing before doing that). It's possible some of the game is remaining in the video memory afterwards. Maybe I need to reboot before starting CUDA processing? That would get to be annoying. Might be better off to just disable GPU processing. ID: 873918 ·

Claggy Volunteer tester Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4	Message 874009 - Posted: 9 Mar 2009, 10:38:39 UTC - in response to Message 873850. Actually, the most current drivers listed for your card are 182.08 http://www.nvidia.com/object/winvista_x86_182.08_whql.html Not sure about the other stuff though. The latest drivers for XP 32bit are: GeForce Driver Release 182 WHQL Claggy ID: 874009 ·

©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.