I see in other questions people are sometimes using this flag, which enables all instruction subsets supported by the local machine. Does it bring a performance improvement? Are there some benchmarks available to measure this?

1 Answer 1


Without running any benchmarks to give you hard numbers, yes, it should make a considerable difference. For example, things like hardware AES support will be picked up when compiling with -march=native (assuming no runtime checks).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.