sai kiran grandhi
sai kiran grandhi

Reputation: 227

Problems in compiling a sample CUDA program

I am new to compiling CUDA C programs on the Linux environment. I was using the Ubuntu 12.04 server.

I want to run the template SDK provide with GPU computing SDK

When I run template by the command

nvcc -o template template.cu

It is returning me template.cu:24: fatal error: cutil_inline.h No such file or directory

Actually the file was in the directory ~/NVIDIA_GPU_Computing_SDK/CUDALibraries/common/inc The same error occurs for the file template_kernel.cu which is the same directory of template

Why this error is coming. Can any one please help me. Is it possible to give the path for these files while launching the template program i.e, giving the include files explicity but not in the program

Upvotes: 0

Views: 200

Answers (1)

Robert Crovella
Robert Crovella

Reputation: 152173

You're having problems because that is not how you build the sample programs.

Based on your description, it appears to you are using a CUDA version prior to CUDA 5.0.

Refer to the file ~/NVIDIA_GPU_Computing_SDK/docs/CUDA_SDK_Release_Notest.txt section "I. (b) Linux Installation Instructions", to learn how to build the sample programs.

As @JackOLantern indicates, you should be using the supplied Makefile

Upvotes: 1

Related Questions