GPU to cpu/system Memory Traffic

Hi all

Firsty, I am a noob in cuda.
I have an Nvidia GTX 1070 and I need to find some kind of computation intensive stress that bogs the pcie Lanes on my system, so some kind of GPU to CPU/ Memory workload .

In essence the purpose of this test is not to stress the GPU but the interconnects and system memory transfer it self

Could someone point me in the right direction

bandwidthTest CUDA sample code