What is the physical level in c plus plus?

Answer 1

When we write code we generally use and write generic routines that can be executed upon any machine. We are not overly concerned with the hardware we are targetting, because abstraction takes care of the machine specifics for us. Thus we can concentrate on writing code specific to the user, rather than worry about the underlying architecture. However, while C++ allows us to use generic and highly abstract interfaces (the standard library is a prime example), it also allows us to use mid-level C-style code, as well as low-level, inline assembly, should we require it. When we use inline assembly we are working at the physical level and coding specifically to the machine, just as if we were writing the code in pure assembler. At this level there is no abstraction whatsoever, other than that provided by the symbolic language itself, and our code will only work on the specific architecture we are targetting. While this allows us to exploit the hardware to the fullest, there are relatively few cases where this is actually necessary. Both C and C++ were specifically designed to produce the machine-specific code for us, and are both capable of optimising that code in order to exploit the underlying hardware's features (that is one of the roles of the compiler). So while we can use C++ to write machine-specific code when we need it, most of the time we will make use of abstraction and generic code that can be run on any machine, and make use of compiler directives to filter any machine-specific code, and only use inline assembly when it is absolutely necessary to do so.