search for: op_rpush

...p them before returning to the dispatch loop. Ok, so I am looking to build a compilation mechanism using LLVM and imagine leveraging optimizations at two distinct levels. I'm sure there must be opportunities to optimize my VM byte code sequences (for example eliminating OP_PUSH and OP_POP or OP_RPUSH and OP_RPOP sequences or reordering stack use in general to be more optimized and things like that). This is obviously something I will need to do myself and probably won't leverage too much of LLVM to help with (although if I'm wrong about this and there is support at this level, I would...