PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

Compiler failed to translate accelerator region
Goto page Previous  1, 2
 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling
View previous topic :: View next topic  
Author Message
mkcolg



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

PostPosted: Fri Jun 21, 2013 9:46 am    Post subject: Reply with quote

What's the compute capability of your device (see the output from the pgaccelinfo utility)?

You may also want to add in error checking after your kernel call to see if the kernel is failing.

Code:
call cal<<<1,np>>>(xi_d,yi_d,ri_d,diff2_d,n,f_d)
istat = cudaThreadSynchronize()
istat = cudaGetLastError()
if (istat.ne.0) then
  print *, cudaGetErrorString(istat)
endif


- Mat
Back to top
View user's profile
camaptrang



Joined: 28 May 2013
Posts: 19

PostPosted: Mon Jun 24, 2013 2:49 am    Post subject: Compiler failed to translate accelerator region Reply with quote

hi Mat,
I did like you tolk. I got "invalid device function". What that mean?
Thank
Back to top
View user's profile
mkcolg



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

PostPosted: Mon Jun 24, 2013 4:25 pm    Post subject: Reply with quote

It's most likey that your device doesn't support double precision. What is your compute capability? (which can be found via the utility pgaccelinfo).

- Mat
Back to top
View user's profile
camaptrang



Joined: 28 May 2013
Posts: 19

PostPosted: Tue Jun 25, 2013 4:53 am    Post subject: Compiler failed to translate accelerator region Reply with quote

Hi Mat,
I use GeForce 310M. Cuda version: 5050
PGI Compiler Option -ta=nvidia,cc12. what is that mean 'cc12'?I must use what decalration for the real number?
thank of avance.
Giang
Back to top
View user's profile
mkcolg



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

PostPosted: Wed Jun 26, 2013 9:42 am    Post subject: Reply with quote

Hi Giang,

Yes, your card doesn't support double precision floating point. This was added to devices with compute capability 1.3 (cc13), but your card is CC 1.2.

For real numbers use "REAL" (without the -r8 flag) or "REAL*4".

Also, make sure you aren't using double precision constant values either (0.0D).

- Mat
Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    PGI User Forum Forum Index -> Programming and Compiling All times are GMT - 7 Hours
Goto page Previous  1, 2
Page 2 of 2

 
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