Performances of reverse_iterator vs iterator?

Question

I am designing a very efficient iterator for a High Performance Computing / Supercomputer application. I was wondering whether the std::reverse_iterator adaptor will introduce some overhead or whether it is likely to compile and achieve the exact same performances as the iterator type it is applied to.

Answer 1

When it comes to performance, don't guess, don't ask, measure . And don't measure with artificial small test programs. Measure speed with your real code, your real application, your real data on real target machines.

C++ as a language does not make any guarantees as to whether std::reverse_iterator introduces overhead or not. The question would be a bit more answerable if you at least told us which compiler and which compiler options you are using, because the performance of standard-library facilities (algorithmic complexity requirements aside) is a quality-of-implementation (QoI) issue. It's unlikely that you would experience a difference in performance with std::reverse_iterator , but who knows?

You also have not told us anything about the underlying iterator that you want to adapt. It may, for instance, be an iterator into a custom container and have different performance characteristics for -- and ++ .