Von -Neumann Architecture vs Harvard Architecture
Before studying Von-Neumann Architecture and Harvard Architecture, must understand computer architecture.
What is computer architecture?
In computer engineering, computer architecture is a set of rules and methods that describe the functionality, organization, and implementation of computer systems. Some definitions of architecture define it as describing the capabilities and programming model of a computer but not a particular implementation.
Von-Neumann Architecture
This model is proposed by Von-Neumann in 1945. A control unit. Arithmetic logic memory unit and inputs/Outputs have consisted of this model. Von- Neumann architecture is based on the stored-program computer concepts. Here program data and instructions are used the same memory for storing the data.
We can see these three main facts where the Von- Neumann Based computer.
1. A single processor is used in Von-Neumann based computer.
2. Both instructions and data are in the same memory
3. The fetch-decode-executive cycle is used for executing the program.
There are three main components in Von-Naumann architecture.
1. Central processing unit
2. Memory
3. Input/Output devices
Harvard Architecture
Most of the computers are followed by Von-Neumann architecture. In that architecture, both instructions and data are stored in the same memory. Same buses are used for fetching data and instructions. That means the CPU cannot do read instructions and read/write data in both functions together. Harvard architecture is the solution to this. In this computer, architecture has been created by separate buses and separate storage for data and instructions. This architecture has been developed to overcome the bottleneck of Von- Neumann architecture. The main advantage of Harvard architecture is CPU can read-write data and access instructions same time because of having separate buses and instructions.
Difference between Von Neumann and Harvard Architecture
Basic of architecture
The Von Neumann architecture model is like a theoretical computer design and this is not a modern computer architecture. This model is based on the concept of stored-program architecture. Here data and programs use the same memory for storing data. John Von Neumann designed this concept in 1945. The Harvard architecture has been created by using the original Harvard Mark I relay-based computer model. Data and instructions use separate buses for storing data.
Memory system
In Von Neumann architecture, there is only one bus system. Thus, the bus used for transferring data and processing instructions. These two operations should be scheduled before because of having one bus and that cannot use separate operations at the same time. But in Harvard Architecture there is a separate memory allocation for transferring data and processing instructions.
Instruction processing
In Von Neumann architecture there should be two clock cycle to complete processing single instruction. In the first cycle, the program counter fetches the instruction from the main memory and in the second clock cycle decode those instructions and then store-related the register in the CPU. But in Harvard architecture, only one clock cycle is enough for processing single instruction.
Cost
When we compare Von Neumann architecture and Harvard architecture Von Neumann architecture is cheaper than Harvard architecture because both data and instructions are used in the same bus system. Therefore, the production cost of lower than Harvard architecture. The cost of Developing a control unit of Harvard architecture is more expensive because of the bus system and complexity.
Uses of architecture
Von Neumann architecture mostly uses for small computers and personal computers like desktop computers. Harvard architecture is not used mostly because the model is still the newest. But microcontrollers and signal processing use this model.
