PGI User Forum
 SearchSearch   MemberlistMemberlist     RegisterRegister   ProfileProfile    Log inLog in 

CUDA-x86.

Parallel for loop: Internal compiler error

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



Joined: 18 Nov 2009
Posts: 3
PostPosted: Mon Nov 23, 2009 5:41 am    Post subject: Parallel for loop: Internal compiler error Reply with quote
Hi,

I'm trying to compile a simple loop where the iteration boundaries are not fix:

Code:

#pragma acc region
{
#pragma acc for
    for (int y = ytl; y < ybr; y++) {
        for (int x = xtl; x < xbr; x++) {
            float v1 = a[x + y*width];
            float v2 = b[x + y*width];
            sum += (v1 - v2);
        }
    }
}


However, the compiler exits with an internal compiler error:
Code:

PGC-F-0000-Internal compiler error. unknown reference      15 (test.c: 46)
PGC/x86-64 Linux 10.0-0: compilation aborted

The loop is inside of a function ending at line number 46.
When I replace xtl by a constant, the compiler generates working code for CUDA. Is there any way to get around this issue?

Best regards,
Richard
Back to top
View user's profile
mkcolg



Joined: 30 Jun 2004
Posts: 4996
Location: The Portland Group Inc.
PostPosted: Mon Nov 23, 2009 1:50 pm    Post subject: Reply with quote
Hi Richard,

I tried to recreate the ICE, but was unable. If you could send an example code which reproduces the error to PGI customer service (trs@pgroup.com), I would appreciate it.

What I did encounter was that I needed to specify the bounds of the a and b arrays in the copyin clause. It's possible that you need to do the same. For example:
Code:
% cat test.c
#include <stdio.h>
#include <accel.h>

float foo (float *a, float *b, int xtl, int ytl, int ybr,int xbr,int width) {
float sum;

#pragma acc region copyin(a[0:xtl], b[0:xtl])
{
    for (int y = ytl; y < ybr; y++) {
        for (int x = xtl; x < xbr; x++) {
            float v1 = a[x + y*width];
            float v2 = b[x + y*width];
            sum += (v1 - v2);
        }
    }
}
return sum;
}
% pgcc -c test.c -Msafeptr -ta=nvidia -V10.0 -Minfo=accel
foo:
      9, Generating copyin(b[:xtl])
         Generating copyin(a[:xtl])
     11, Loop is parallelizable
         Accelerator kernel generated
         11, #pragma acc for parallel, vector(256)
         15, Sum reduction generated for sum
     12, Loop is parallelizable


Hope this helps,
Mat
Back to top
View user's profile
hwswcodesign



Joined: 18 Nov 2009
Posts: 3
PostPosted: Tue Nov 24, 2009 7:29 am    Post subject: Reply with quote
Hi Mat,

using the copyin clause was the right hint!
I had still some problems since I used structs to store some parameters and got the same error message. Using normal variables solved the problem!

Richard
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 © 2001, 2002 phpBB Group