What are the merits of volatile in c plus plus?

Answer 1

The volatile specifier informs the compiler that an object could be modified by something that is external to the thread of execution; something that is not part of the program. For instance, when reading a hardware clock, you would rightly expect the value of that clock to change from one read to the next. But from the compiler's perspective, the clock remains constant because your code does not alter it between reads and is unaware of any external code that controls its value. Therefore the compiler will (rightly) optimise away what it believes to be redundant read operations which could easily break the semantics of your code.

By declaring an object volatile (even a constant object), you inform the compiler that the object may change between successive read or write operations and that all redundancy optimisations must not be applied to any code that uses that object.

Although often used as a method of synchronisation between concurrent threads within the same thread of execution that share the same object, you must not use it for this purpose. C++ provides very specific mechanisms for this purpose, such as an atomic, a mutex or a condition_variable. The volatile specifier must only be used when writing low-level code that deals directly with hardware, such as when writing device driver and embedded system software.

Answer 2

None. It simply warns to compiler that the variable in question is target of asynchronous access.