I’m trying to add vector instructions to my target. Does anybody know if LLVM has an option to enable vector instructions? In other words if LLVM sees a possible optimization where it could use a vector instruction it would actually use it.
Any help is appreciated.