Von Neumann Architecture
The important parts of this architecture include the following:
- Processing unit: It contains an Arithmetic Logic Unit (ALU) and a set of working registers for the processor.
- Control unit: It encompasses control mechanism to carry out functions and includes instruction register and program counter.
- Memory: It stores data and instructions used for applications and system processing, addresses mapping for external mass storage, and I/O mechanisms.
The traditional Von Neumann architecture describes a stored-program computer, which does
not allow fetching of instructions and data operations to occur at the same time. The reason
is that the architecture uses a commonly shared bus to access both. This has led to limiting
the performance of the system. The structure of a typical Von Neumann machine is shown
Harvard Computer Architecture
The characteristics of Harvard Computer architecture are as follows:
- The Harvard architecture is a stored-program computer system, which has separate sets of addresses and data buses to read and write data to the memory and also for fetching instructions.
- Basically, the Harvard architecture has physically distinct storage and signal paths access to data and instructions, respectively.
- In the Harvard architecture, two memories need not share characteristics.
- The structure and width of the word, timing characteristics, mechanism of implementation, and structure of addresses can vary, while program instructions reside in read-only memory, the program data often needs read-write memory.
- The requirements of instruction memory can be larger due to the fact that some systems have much more instruction memory than data memory. Therefore, the width of instruction addresses happens to be wider than that of data addresses.