Floating point exception

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello.

This exception appears every time when I run the examples from SDK and trying to link CUDA libraries with my application.

Environment:

Tesla c2070;

GNU/Linux (kernel 2.4.32);

gcc 4.3.3;

binutils 2.22;

glibc 2.3.6;

gpucomputingsdk_4.1.28_linux;

At first I thought that the problem lies in the different fields hash values ​​of the header elf-files. But after bringing them to the same value of the problem has not disappeared.

For example:

test.c

#include <stdio.h>

void main() {

  printf("test");

}

linux# gcc test.c -o test Wl,--hash-style=both -I/includes -L/lib -lcudart

linux# ./test

Floating point exception.

/Thanx