PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

compiler errors

 
Post new topic   Reply to topic    PGI User Forum Forum Index -> Accelerator Programming
View previous topic :: View next topic  
Author Message
spam.me



Joined: 25 Jun 2009
Posts: 6

PostPosted: Thu Oct 08, 2009 3:48 am    Post subject: compiler errors Reply with quote

Hello,

If I want to compile the following loop section
Code:

#pragma acc region
{
[...]

#pragma acc for parallel private(ip,ir)
   FORALL 
      dtw_0[ip][ir] +=    0.5*hval[ir]*(visc_w_2d[ip  ][ir+1] - visc_w_2d[ip  ][ir-1])*0.5*hval[ir]*(w_0[ip  ][ir+1] - w_0[ip  ][ir-1]);

[...]
}


pgcc tells me

Code:

pgcc -mp  -O3 -Mnoopenmp -ta=nvidia -Minfo=accel -fast -Msafeptr  ...

[...]

PGC-F-0000-Internal compiler error. Generating a negative offset into dimension of target array       1 (c/esel.c: 1383)
PGC/x86-64 Linux 9.0-4: compilation aborted


But pgcc compiles fine without -ta=nvidia -Minfo=accel, I tried also #pragma acc for host without an effect. What are my options?

Secondly, I would ask why pgcc doesn't show the correct line number of the source code where the problem was detected?

Thank you!

With kind regards,


Futher information
Code:

#define FORALL  for(ip = 0; ip <  ny; ip++) for(ir = 0; ir <  nx; ir++)

[...]

/* allocate the matrix */
  double      **field;

  sizey = (size_t)(ny  + 2*offy);
  sizex = (size_t)(nx  + 2*offx );

  field       = (double **)CALLOC(sizey,      sizeof(double *))+offy;
  field[-offy] = (double  *)CALLOC(sizex*sizey,sizeof(double)  )+offx;

  for(i = 1-offy;i<ny+offy;i++) field[i]=field[i-1]+sizex;

  return field;

Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 5952
Location: The Portland Group Inc.

PostPosted: Thu Oct 08, 2009 9:46 am    Post subject: Reply with quote

Hi Spam.me,

Can you please send a report to PGI customer service and include an example of the error? Internal compiler errors are basically assertions in the compiler itself and occur when the compiler enters an unexpected state. We'll need to have one of our engineers look at the issue to determine what we're doing wrong.

Thanks,
Mat
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
Page 1 of 1

 
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