Question

My matrix addition example:

 __global__ void matrix_add(float *a, float*b, float *c, int N)
{
    int index;
    int Row = blockIdx.y * blockDim.y + threadIdx.y;
    int Col = blockIdx.x * blockDim.x + threadIdx.x;

    int index = Row * N + Col;      // N is the order of the square matrix

    cd[index]= ad[index] + bd[index];

}

Can I use printf or any other similar function in above kernel? So that I won’t need to transfer data from device to host memory (i.e. cudaMemcpyDeviceToHost). If yes then how? If no then why not?

Answer 1

You could use printf(..) but only for cc2.x or higher.
You could read more about this in the CUDA programming guide Appendix B.16.