PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

Trouble Getting Started CUDA/PGI Fortran
Goto page 1, 2, 3, 4, 5, 6, 7  Next
 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Accelerator Programming
View previous topic :: View next topic  
Author Message
PSchneid2000



Joined: 22 Jun 2010
Posts: 1

PostPosted: Tue Feb 15, 2011 10:13 am    Post subject: Trouble Getting Started CUDA/PGI Fortran Reply with quote

Using:
Windows 7 Ultimate 64-bit
Tesla C2050
nVidia Driver 260.81
CUDA Toolkit 3.2 (32 and 64 bit installed)
MS Visual Studio 2010
PGI Fortran 11.2

Code: (compiled in 32-bits)
PROGRAM PRUEBA
USE CUDAFOR
IMPLICIT NONE

INTEGER :: NUMBX
INTEGER, DEVICE :: NUMBX_con


NUMBX = 320
!
NUMBX_con = 2


END PROGRAM


Debug Output:
'PVFProject6_CUDA.exe': Loaded 'C:\Windows\SysWOW64\ntdll.dll', No symbols loaded.
'PVFProject6_CUDA.exe': Loaded 'C:\Windows\syswow64\kernel32.dll', No symbols loaded.
'PVFProject6_CUDA.exe': Loaded 'C:\Windows\syswow64\KERNELBASE.dll', No symbols loaded.
'PVFProject6_CUDA.exe': Loaded 'c:\program files (x86)\pgi\win32\2011\cuda\3.2\bin\cudart32_32_16.dll', No symbols loaded.
The thread '0.0' (0x918) has exited with code 0 (0x0).
The program 'PVFProject6_CUDA.exe: PGI Debug Engine' has exited with code 0 (0x0).

As far as I can tell, all my path and variable statements are correct. Can someone give me an idea to start looking to find the problem?
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6134
Location: The Portland Group Inc.

PostPosted: Tue Feb 15, 2011 3:09 pm    Post subject: Reply with quote

Hi PSchneid2000,

Exit code 0 usually means your program ran successfully. Are you seeing a different error?

- Mat
Back to top
View user's profile
Dolf



Joined: 22 Mar 2012
Posts: 105

PostPosted: Mon Oct 29, 2012 10:40 am    Post subject: RE: Reply with quote

Hi Mat and everyone,

I am having the same problem actually.
When I planted many breakpoints all over the code and run it in debug mode, I was able to make it run, till I get this message in VS 2010:

No source available
Call stack location:
pgf90_dev_alloc04() Line 348 in "dev_allo.c" address: 14006bd20

the debugger could not locate the source file 'src\dev_allo.c'.

what does that mean?

also, if I run the code in release mode, I get an error saying that memcopy error, which means that one of the kernel subroutines have problems and cannot copy from device to host of the opposite.

please help.

Dolf
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 6134
Location: The Portland Group Inc.

PostPosted: Mon Oct 29, 2012 11:14 am    Post subject: Reply with quote

Quote:
what does that mean?
It just means that the debugger is attempting to step into code which wasn't compiled with debugging information enabled. In this particular cases, "pgf90_dev_alloc04" is part of the PGI run time libraries so you wouldn't have access to the source.

Quote:
if I run the code in release mode, I get an error saying that memcopy error, which means that one of the kernel subroutines have problems and cannot copy from device to host of the opposite.
Most likely when optimization is applied, something changes in your code which causes the problem. You can try debugging with optimization enabled to try and narrow down the problem.

- Mat
Back to top
View user's profile
Dolf



Joined: 22 Mar 2012
Posts: 105

PostPosted: Mon Oct 29, 2012 4:26 pm    Post subject: RE: Reply with quote

Hi Mat,

Unfortunately, the optimization did not work either.
I have copied the error message when I run in release mode:



COMPUTING PRESSURE FIELD ON THE CURRENT GRID...

starting grid level 5
GRV qni kernel error
unknown error

GRV qnj kernel error
unknown error

istat = 30
0: copyout Memcpy (host=0x1dd0f788, dev=0x4121c5e8, size=8) FAILED: 30(unknown e
rror)
Press any key to continue . . .

the most strange thing is, istat (error code after calling the kernel) is 30!
I thought the error code is from 0, to 11 only.

here is how I call the kernel, how can I check for errors in case I encounter any:

grid = dim3(ceiling(real(nx-1)/threads%x), &
ceiling(real(ny)/threads%y),1)

call GetReynVarqnj_kernel<<<grid,threads>>>(nx,ny,ndx,ndy, &
p,hnew,hjmin,hjmax,cohjmx,s,l,kd,zdatLowDev, &
qndatLowDev,zdatMidDev,qndatMidDev,zdatHighDev,qndatHighDev,qnj)

istat = cudaGetLastError() <=== equals 30!

please advice what could be the meaning of such GetLastError.

thanks,
Dolf
Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    PGI User Forum Forum Index -> Accelerator Programming All times are GMT - 7 Hours
Goto page 1, 2, 3, 4, 5, 6, 7  Next
Page 1 of 7

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © phpBB Group