In general, the machine is a digital information can be found in random locations (cells) and storage is a problem, how to find her. This problem can be solved in different ways. The most common approach, in which all the cells in memory are numbered, and the required of them were wanted on a number called the address. Memory, using this principle, known as targeted. Sequential memory can also be converted to address if the carrier next to the symbol store by its address.
The serial nature of memory inevitably reduces performance. Indeed, if after the symbol, which is located in one "end" of the memory, information is needed from the other "end", you need a useless review the contents of all the intermediate cells.
Targeted memory with the same time access to all the cells is called random-access memory (RAM). Access time to a cell with any address in the most advanced of these is tenths and even hundredths of a microsecond! Performance as operating devices, which are based combinational circuits, may be even higher.
What if the program itself will not do in the CU from external media, but from the RAM?
It is believed that this idea belongs to the American Hungarian-born John von Neumann, who took very active part in the development of computer technology. Method, in which the coded program of the CA is located along with the processed information in the memory of CA, was called the principle of in-memory programs, and self RAM - random access memory (RAM). Its use, he found already in early models of computer and since then has been one of the key, integral features of any computer.
Targeted memory leads to the fact that the team, which theoretically splits algorithm for solving any problem, in general, contain the following information:
1) The address of the transformed character;
2) an indication of the unit operation to be perform on this character;
3) address the data memory where you want to put the result;
4) The address data memory which stores the following command program.