Is there any approach of shared mutex for a short shared-operations and few unique operations?

As known std::shared_timed_mutex (C++14) and std::shared_mutex (C++17) has a performance advantage over std::mutex only for a very long std::shared_locks (reads-operations), and when std::shared_locks more than std::unique_locks.

But even if most of its are read operations, and they are short, then std::mutex will be faster - that is why in the C++11 standard std::shared_mutex and std::shared_timed_mutex were not included immediately.

Is there any approach or implementation of some kind of shared mutex, which has a performance advantage over std::mutex, also for a short operations, if the reads are not just more, and a lot more as below?

std::unique_lock / std::shared_lock = 1 / 1 000 000 000 000

Is there any approach of shared mutex for a short shared-operations and few unique operations?