Common compiling errors in CUDA?

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi!
Fedora 10
gcc 4.3.2
Nvidia driver 190.18
CUDA 2.3

The CUDA Toolkit installation defaults to /usr/local/cuda
The CUDA SDK installation defaults to /root/NVIDIA_GPU_Computing_SDK

already done:

export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib:$LD_LIBRARY_PATH

and have placed them in ~/.bash_profile.

now I have error:
1, make in /NVIDIA_GPU_Computing_SDK/C, errors below:

/usr/local/cuda/lib/libcudart.so: undefined reference to __cxa_get_exception_ptr@CXXABI_1.3.1' collect2: ld returned 1 exit status make[1]: *** [../../bin/linux/release/sortingNetworks] Error 1 make[1]: Leaving directory /home/ROLY/NVIDIA_GPU_Computing_SDK/C/src/sortingNetworks’
make: *** [src/sortingNetworks/Makefile.ph_build] Error 2

2, make in /NVIDIA_GPU_Computing_SDK/C/src/deviceQuery, errors below:

/usr/local/cuda/lib/libcudart.so: undefined reference to `__cxa_get_exception_ptr@CXXABI_1.3.1’
collect2: ld returned 1 exit status
make: *** […/…/bin/linux/release/matrixMul] Error 1

I found that some others faced the same problem last year, they regarded it as “gcc’s version problem”,
but I tested in gcc4.3.2 and gcc3.4.6, the same errors. What should I do?

Cheers!

2

Are you sure that you are linking with the correct version of libstdc++?