Do Not Use Depth Pre-pass

Depth pre-pass is redundant on deferred rendering architectures.

On graphics hardware that employs a deferred rendering architecture such as PowerVR, an application should not perform a depth pre-pass as there is no performance benefit. Performing this operation would be a waste of clock cycles and memory bandwidth. This is because the hardware will detect and remove occluded (opaque) geometry from the pipeline automatically during rasterization, before fragment processing begins.