CUDA - invalid device function


Advanced search

Questions and Answers : GPU applications : CUDA - invalid device function
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928007 - Posted: 22 Aug 2009, 15:10:50 UTC
Last modified: 22 Aug 2009, 15:24:08 UTC

Please, look at this task:

http://setiathome.berkeley.edu/result.php?resultid=1339584737

setiathome_CUDA: Found 1 CUDA device(s):
   Device 1 : GeForce 8500 GT 
           totalGlobalMem = 536543232 
           sharedMemPerBlock = 16384 
           regsPerBlock = 8192 
           warpSize = 32 
           memPitch = 262144 
           maxThreadsPerBlock = 512 
           clockRate = 918000 
           totalConstMem = 65536 
           major = 1 
           minor = 1 
           textureAlignment = 256 
           deviceOverlap = 1 
           multiProcessorCount = 16 
setiathome_CUDA: CUDA Device 1 specified, checking...
   Device 1: GeForce 8500 GT is okay
SETI@home using CUDA accelerated device GeForce 8500 GT
V12 modification by Raistmer 
Priority of worker thread rised successfully
Priority of process adjusted successfully
Total GPU memory 0	 free GPU memory 0
Cuda error 'cufftPlan1d(&fft_analysis_plans[FftNum], FftLen, CUFFT_C2C, NumDataPoints / FftLen)' in file 'd:/BoincSeti_Prog/sinbad_repositories/LunaticsUnited/SETI_CUDA_MB_exp/client/cuda/cudaAcc_fft.cu' in line 101 : invalid device function .
setiathome_CUDA: CUDA runtime ERROR in plan FFT. Falling back to HOST CPU processing...


Task's end after .... Run time 26902.59!!!

Computer is GenuineIntel
Intel(R) Core(TM)2 Quad CPU Q6700 @ 2.66GHz [x86 Family 6 Model 15 Stepping 11]
(4 processors) with NVIDIA GeForce 8500 GT (511MB) driver: 19038
Optimized application by Win32 Lunatics' Unified Installer v0.2 (working well with other geforce CUDA).

Bye,
Francesco

PS: posted on lunatic forum too.
ID: 928007 · Report as offensive
Profile Gundolf Jahn

Send message
Joined: 19 Sep 00
Posts: 3184
Credit: 446,358
RAC: 0
Germany
Message 928022 - Posted: 22 Aug 2009, 16:09:47 UTC - in response to Message 928007.  
Last modified: 22 Aug 2009, 16:10:42 UTC

setiathome_CUDA: Found 1 CUDA device(s):
   Device 1 : GeForce 8500 GT 
           totalGlobalMem = 536543232 
           sharedMemPerBlock = 16384 
           regsPerBlock = 8192 
...
Total GPU memory 0	 free GPU memory 0
...
Falling back to HOST CPU processing...


Task's end after .... Run time 26902.59!!!

No wonder that it took so long, it was run on the CPU due to lack of available memory.

Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)

SETI@home classic workunits 3,758
SETI@home classic CPU time 66,520 hours
ID: 928022 · Report as offensive
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928027 - Posted: 22 Aug 2009, 16:20:22 UTC - in response to Message 928022.  


...
Total GPU memory 0 free GPU memory 0
...
Falling back to HOST CPU processing...

Task's end after .... Run time 26902.59!!!

No wonder that it took so long, it was run on the CPU due to lack of available memory.

Gruß,
Gundolf

Ok, but why?
May be in that PC I'm using all memory with a big screen resolution?

I see:
totalGlobalMem = 536543232
...
Total GPU memory 0 free GPU memory 0

On the task working on other CUDA I see that totalGlobalMem = Total GPU memory and that free GPU memory is < of Total GPU memory.

What may happen?

Francesco

ID: 928027 · Report as offensive
Profile Gundolf Jahn

Send message
Joined: 19 Sep 00
Posts: 3184
Credit: 446,358
RAC: 0
Germany
Message 928035 - Posted: 22 Aug 2009, 16:29:49 UTC - in response to Message 928027.  
Last modified: 22 Aug 2009, 16:30:53 UTC

Did you try a reboot (full power cycle)? Something might be stuck in graphics memory.
ID: 928035 · Report as offensive
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928043 - Posted: 22 Aug 2009, 16:59:17 UTC - in response to Message 928035.  

Yes.

Maybe the problem is like this one: http://setiathome.berkeley.edu/forum_thread.php?id=54788&nowrap=true#921096
But I have all new soft & drivers.

Bye
Francesco
ID: 928043 · Report as offensive
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928149 - Posted: 23 Aug 2009, 7:14:24 UTC - in response to Message 928035.  

Did you try a reboot (full power cycle)? Something might be stuck in graphics memory.


after restart I see:

WU true angle range is : 0.444235
After app init: total GPU memory 536543232 free GPU memory 480770560
setiathome_CUDA: Found 1 CUDA device(s):
Device 1 : GeForce 8500 GT
totalGlobalMem = 536543232
sharedMemPerBlock = 16384
regsPerBlock = 8192
warpSize = 32
memPitch = 262144
maxThreadsPerBlock = 512
clockRate = 918000
totalConstMem = 65536
major = 1
minor = 1
textureAlignment = 256
deviceOverlap = 1
multiProcessorCount = 16
setiathome_CUDA: CUDA Device 1 specified, checking...
Device 1: GeForce 8500 GT is okay
SETI@home using CUDA accelerated device GeForce 8500 GT
V12 modification by Raistmer
Priority of worker thread rised successfully
Priority of process adjusted successfully
Total GPU memory 0 free GPU memory 0
Cuda error 'cufftPlan1d(&fft_analysis_plans[FftNum], FftLen, CUFFT_C2C, NumDataPoints / FftLen)' in file 'd:/BoincSeti_Prog/sinbad_repositories/LunaticsUnited/SETI_CUDA_MB_exp/client/cuda/cudaAcc_fft.cu' in line 101 : invalid device function .
setiathome_CUDA: CUDA runtime ERROR in plan FFT. Falling back to HOST CPU processing...

ID: 928149 · Report as offensive
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928422 - Posted: 24 Aug 2009, 16:20:03 UTC

More info

I have made a full reinstall of Win32 Lunatics' Unified Installer v0.2
and Cuda2.3rt_fft_dlls.7z

As BOINC 6.6.36 stats I read:

24.08.2009 14:53:48             Starting BOINC client version 6.6.36 for windows_intelx86
24.08.2009 14:53:48             log flags: task, file_xfer, sched_ops
24.08.2009 14:53:48             Libraries: libcurl/7.19.4 OpenSSL/0.9.8j zlib/1.2.3
24.08.2009 14:53:48             Data directory: D:\Boinc
24.08.2009 14:53:48             Running under account Luca
24.08.2009 14:53:48             Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad CPU    Q6700  @ 2.66GHz [x86 Family 6 Model 15 Stepping 11]
24.08.2009 14:53:48             Processor features: fpu tsc pae nx sse sse2 mmx
24.08.2009 14:53:48             OS: Microsoft Windows XP: Professional x86 Edition, Service Pack 3, (05.01.2600.00)
24.08.2009 14:53:48             Memory: 2.00 GB physical, 3.85 GB virtual
24.08.2009 14:53:48             Disk: 270.44 GB total, 268.94 GB free
24.08.2009 14:53:48             Local time is UTC +2 hours
24.08.2009 14:53:48             CUDA device: GeForce 8500 GT (driver version 19038, compute capability 1.1, 512MB, est. 5GFLOPS)
24.08.2009 14:53:48     SETI@home       Found app_info.xml; using anonymous platform
24.08.2009 14:53:48             [error] Deleting file MB_6.08_CUDA_V12_VLARKill_FPLim2048.exe while in use
24.08.2009 14:53:48             [error] Deleting file cudart.dll while in use
24.08.2009 14:53:48             [error] Deleting file cufft.dll while in use
24.08.2009 14:53:48             [error] Deleting file libfftw3f-3-1-1a_upx.dll while in use
....


Do you know the meaning of the errors?

Now I see 5 task running (4 CPU and 1 GPU) but I have to wait the end of the GPU's task to look in the task result.

Bye,
Francesco
ID: 928422 · Report as offensive
Profile Francesco Forti
Avatar

Send message
Joined: 24 May 00
Posts: 334
Credit: 204,421,005
RAC: 15
Switzerland
Message 928616 - Posted: 25 Aug 2009, 14:26:16 UTC - in response to Message 928422.  


Now I see 5 task running (4 CPU and 1 GPU) but I have to wait the end of the GPU's task to look in the task result.


All OK now.
Problem closed
FF
ID: 928616 · Report as offensive

Questions and Answers : GPU applications : CUDA - invalid device function


 
©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.