i++ and ++i are bytecode-equivalent (Java)Idk if you've heard of performance differences between post-increment (e.g. i++) and pre-increment (e.g. ++i). The idea is that post-increment needs to be implemented as tmp = i ; i = i + 1; return tmp whereas pre-increment can be implemented as i = i + 1; return i and so is more efficient. But I was thinking about the following question: In Java, is there a performance difference between i++ and ++i if the result is never used? How can I be sure that i++ [return value unused] is always executed as just i = i + 1? I found this stackoverflow answer that uses the following neat technique: Just write two programs that are identical except for i++ vs ++i, and observe that the java bytecode produced (.class file) is exactly the same. So no semantic understanding of the bytecode is required, just (classfile1 diff classfile2 or md5sum classfile1 ; md5sum classfile2)