What is the difference between static and dynamic memory?

Answer 1

There is no such thing as dynamic memory, but we often use the term to mean non-static memory. Static memory is where all constant variables, global variables and static variables are allocated in static memory. Static memory is a contiguous block of memory of fixed length divided into two sections: initialised memory and uninitialised memory. Initialised variables and constants are allocated to the initialised section and are stored as a bitmap image in the executable itself. Uninitialised variables are allocated in the uninitialised section which is typically a block of zero-initialised memory. No bitmap image is required for uninitialised memory, only its length need be stored in the executable. The compiler can determine the amount of memory required for each constant and variable.

Non-static memory is essentially all other memory used by a program. Every thread has its own call stack allocated as each thread is instantiated and released as each thread is destroyed. All stacks are fixed-length (determined at design time). The call stack is used to store local variables, automatic variables and temporary variables, as well as to implement exception handling.

The heap (or free store) is essentially all other memory available to the program and is used to instantiate all anonymous variables. The heap is what we generally mean by dynamic memory because anonymous variables are created and destroyed as and when they are required at runtime. Since we cannot refer to an anonymous variable by name (being anonymous means they have no name), we must refer to them solely by their address. This means we must use pointer variables to store those addresses. If we have a large number of anonymous variables of the same type to keep track of, then we typically allocate them in contiguous memory and maintain a single pointer to the start of the allocation. This single pointer allows us to access the memory just as we would a built-in array. We also need to keep track of how many variables (elements) are allocated in the array to avoid buffer overruns. This allows us to easily allocate and release any number of anonymous variables at runtime.

However, in C++ we rarely use "dynamic memory" directly. Although necessary in C Programming, C++ programmers prefer to use resource handles or "smart" pointers instead. In this way the language itself handles the underlying memory allocations for us, ensuring that memory is allocated and released in a timely manner, and thus minimising the risk of creating resource leaks -- a common problem with low-level C Programming. The C++ std::vector is an example of a resource handle and most non-trivial C++ programs will include one or more std::vector<std::unique_ptr<T>> variables to keep track of any number of pointers to type T objects.

Answer 2

There is no such thing as dynamic memory, but we often use the term to mean non-static memory. Static memory is where all constant variables, global variables and static variables are allocated in static memory. Static memory is a contiguous block of memory of fixed length divided into two sections: initialised memory and uninitialised memory. Initialised variables and constants are allocated to the initialised section and are stored as a bitmap image in the executable itself. Uninitialised variables are allocated in the uninitialised section which is typically a block of zero-initialised memory. No bitmap image is required for uninitialised memory, only its length need be stored in the executable. The compiler can determine the amount of memory required for each constant and variable.

Non-static memory is essentially all other memory used by a program. Every thread has its own call stack allocated as each thread is instantiated and released as each thread is destroyed. All stacks are fixed-length (determined at design time). The call stack is used to store local variables, automatic variables and temporary variables, as well as to implement exception handling.

The heap (or free store) is essentially all other memory available to the program and is used to instantiate all anonymous variables. The heap is what we generally mean by dynamic memory because anonymous variables are created and destroyed as and when they are required at runtime. Since we cannot refer to an anonymous variable by name (being anonymous means they have no name), we must refer to them solely by their address. This means we must use pointer variables to store those addresses. If we have a large number of anonymous variables of the same type to keep track of, then we typically allocate them in contiguous memory and maintain a single pointer to the start of the allocation. This single pointer allows us to access the memory just as we would a built-in array. We also need to keep track of how many variables (elements) are allocated in the array to avoid buffer overruns. This allows us to easily allocate and release any number of anonymous variables at runtime.

However, in C++ we rarely use "dynamic memory" directly. Although necessary in C programming, C++ programmers prefer to use resource handles or "smart" pointers instead. In this way the language itself handles the underlying memory allocations for us, ensuring that memory is allocated and released in a timely manner, and thus minimising the risk of creating resource leaks -- a common problem with low-level C programming. The C++ std::vector is an example of a resource handle and most non-trivial C++ programs will include one or more std::vector<std::unique_ptr<T>> variables to keep track of any number of pointers to type T objects.