Data Layout in Stack (Optimization)

Does LLVM do any data layout in the stack to optimize cache line use or prefetch ? I.e. Is the ordering of my calls to CreateAlloca rearranged?

If yes, is there some doc on it?