cudaMemcpyDevicetoHost

Hi

I have following statement in my program:

cudaMemcpy(c, dev_c, N * sizeof(int), cudaMemcpyDevicetoHost);

and I am getting the following error:

compaq@ubuntu:~/CUDA/Programs$ nvcc example5.cu
example5.cu(35): error: identifier “cudaMemcpyDevicetoHost” is undefined

1 error detected in the compilation of “/tmp/tmpxft_00000a1b_00000000-4_example5.cpp1.ii”.

However, I donot have any problem with the following statement:

cudaMemcpy(dev_b, b, N * sizeof(int), cudaMemcpyHostToDevice);

Any comment.

Thanks,

Wei

There is a difference between

cudaMemcpyDevicetoHost

and

cudaMemcpyHostToDevice

I will leave it to the reader to find it (keeping in mind that C is a case sensitive language)…