Hi,
I’m trying to compile the CUDA toolkit’s examples with Fedora 14 x86_64
I’ve installed the FW 260 via RPMFusion, GCC 4.5.1, the CUDA toolkit (for Fedora 13, as root), the gpucomputingSDK ( as user ), etc…
but I got this error when I try to make the “C” examples of the gpucomputingSDK:
Of couse, I’ve added the corresponding
/usr/local/cuda/lib64
/usr/local/cuda/lib
to the /etc/ld.so.conf and I’ve performed a ldconfig as root.
The FW 260 are correctly installed, OpenGL/glexgears works perfectly and the OpenCL ICD is present at /etc/OpenCL/vendors/nvidia.icd ( pointing to libcuda.so ).
I’m not sure why the examples cannot find the OpenCL library, because it’s present in /usr/lib64/nvidia/libOpenCL.so.1 and also in
/usr/lib64/nvidia/libOpenCL.so.1.0.0
Btw, I’ve correctly installed it in both Ubuntu 10.10 and OpenSUSE 11.3, so really cannot figure what’s the problem.
Seems the symbolink library links are not updated or something similar… Any idea, pls?
thx
Hi,
I’m trying to compile the CUDA toolkit’s examples with Fedora 14 x86_64
I’ve installed the FW 260 via RPMFusion, GCC 4.5.1, the CUDA toolkit (for Fedora 13, as root), the gpucomputingSDK ( as user ), etc…
but I got this error when I try to make the “C” examples of the gpucomputingSDK:
Of couse, I’ve added the corresponding
/usr/local/cuda/lib64
/usr/local/cuda/lib
to the /etc/ld.so.conf and I’ve performed a ldconfig as root.
The FW 260 are correctly installed, OpenGL/glexgears works perfectly and the OpenCL ICD is present at /etc/OpenCL/vendors/nvidia.icd ( pointing to libcuda.so ).
I’m not sure why the examples cannot find the OpenCL library, because it’s present in /usr/lib64/nvidia/libOpenCL.so.1 and also in
/usr/lib64/nvidia/libOpenCL.so.1.0.0
Btw, I’ve correctly installed it in both Ubuntu 10.10 and OpenSUSE 11.3, so really cannot figure what’s the problem.
Seems the symbolink library links are not updated or something similar… Any idea, pls?
thx