One of the best ways to understand how the CLR works internally is to have a look at the SSCLI [a.k.a Rotor] source code. In How the SSCLI [a.k.a. Rotor] managed "new" works, I have used Rotor source code to discuss what happens behind the scenes when you use the managed new operator to instantiate a managed type, discussing implications of object size, how requests from multiple threads are handled, the different validations that are done, when out-of-memory exception is thrown, etc. If you are...