How to enable coalesced memory access for this cuda kernel Optimizing a simple ray tracer using cuda

this post is closed.