CUDA shared memory issue in outputs depending on extern declaration and size of array

Question

If I am experimenting with shared memory in CUDA and I do not understand its behaviour in this bit of code. I have a pretty basic kernel:

__global__ void sum         


        

Answer 1

If you use the extern qualifier you need to pass the size of the shared memory when launching the kernel.
kernel<<< blocks, threads, size>>>(...)
The size parameter is the size of shared memory in Bytes.