The goal of this exercise is to understand better the size of C++ objects in memory, to understand some differences between 32bit and 64bit and to learn to use some basic tools.

Examine the size of basic data types in memory, int, long, float, double and pointers, for example. Hint: use the sizeof() function.
Examine the size of the following class in memory (not the generated object file size) after compiling for both 32bit and 64bit:
```
class foo
{
  double a;
  bool agood;
  double b;
  bool bgood;
  double c;
  bool cgood;
  double d;
  bool dgood;
};
```
Do you understand any difference that you see?
Write a little program to create a very large vector of such objects. Add something to the program, for example “std::cin >> i”, such that it subsequently stops waiting for input in order that you can examine the process with command line tools like top or pmap from a separate shell, etc. What do you see? Try it for both 32bit and 64bit.
Do you see how to reduce the memory usage (for both 32bit and 64bit) by modifications to the class?
What happens if the class has a virtual function? (Again, for both 32bit and 64bit.)
Write a small program that makes a very large number (i.e. millions) of small memory allocations (try 4 bytes, class foo above, 100 bytes in subsequent runs). Again stop the program each time with some trick like waiting for input such that you can examine it with top, pmap, etc. Do you understand what you see? Try this test for both 32bit and 64bit.

Data Sizes

References