An easy method for 3D array in kernels

Accelerated Computing CUDA CUDA Programming and Performance
1

If you have a 3D for loop to translate into a kernel

for ( i = 1; i < 5 ; i++){
for ( j = 1; j < 5 ; j++){
for ( k = 1; k < 5 ; k++){
a[i][j][k]= b[i][j][k]+c[i][j][k);
}}}

Define your kernel parameters as follows:

    dim3 grid (5);
    dim3 block (5,5);

    kernel<<< grid,block >>>();

And your kernel

    kernel()
    {
    ix = blockIdx.x + 1;
    iy = threadIdx.y + 1;
    iz = threadIdx.x + 1;

    a[ix][iy][iz]=b[ix][iy][iz] +c[ix][iy][iz];

    }