CUDA SDK compile make Function Pointers problem cannot complete the compile part

Hi,

This is my first post and I am hoping a few kind souls can help me out. I have been searching the forums for a solution and I was unsuccessful. I am trying to compile SDK 3.2.16 by typing “make.”

relevant specs about my comp:

Fedora 14 64bit, cuda tool kit 3.2.16, gcc 4.5.1, NVIDIA SDK 3.2.16, and my current gpu is a GTS 250.

This is my current error code:

[root@pepper C]# pwd

/home/ummm/NVIDIA_GPU_Computing_SDK/C

[root@pepper C]# make

make[1]: Entering directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Leaving directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Entering directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Leaving directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Entering directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Leaving directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/common'

make[1]: Entering directory `/home/ummm/NVIDIA_GPU_Computing_SDK/shared'

make[1]: Leaving directory `/home/ummm/NVIDIA_GPU_Computing_SDK/shared'

make -C src/FunctionPointers/ 

make[1]: Entering directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/src/FunctionPointers'

FunctionPointers_kernels.cu(62): error: calling a host function from a __device__/__global__ function is not allowed

FunctionPointers_kernels.cu(62): error: calling a host function from a __device__/__global__ function is not allowed

2 errors detected in the compilation of "/tmp/tmpxft_00004f0a_00000000-4_FunctionPointers_kernels.cpp1.ii".

make[1]: *** [obj/x86_64/release/FunctionPointers_kernels.cu_20.o] Error 2

make[1]: Leaving directory `/home/ummm/NVIDIA_GPU_Computing_SDK/C/src/FunctionPointers'

make: *** [src/FunctionPointers/Makefile.ph_build] Error 2

I am willing to provide any other information, Any help would be much appreciated. Thank you in advance.

Just skip the function pointers example, it cannot be compile to run for your GPU.

Thanks for responding to my issue. I have a few further questions if you don’t mind me asking. First, how would I go about skipping the function pointer example? And secondly, when you mention that “it cannot be compile to run for your gpu,” do you mean the GTS 250 does not support it?

My GTX560 ti is being delivered on Tuesday, so I was hoping I can get the CUDA installed and ready. Should I wait until I receive my GTX560 before continuing?