Types of processor
Von Neumann architecture:
Specifies the basic components of the computer and processor in which a shared memory and bus is
used for data and instructions
- The stored program concept can be defined as follows, machine code and instructions are
fetched and executed serially by a processor that performs arithmetic and logical operations
- The instructions are fetched sequentially and stored in the main memory, the sequence can
only change by a jump instruction
- For this to happen, the program must be within the main memory
- The instructions are fetched, decoded and executed by the processor one at a time
- This means one bus is for transferring data and instructions, this can lead to collisions and
bottlenecks in the computer. This is done as it simplifies the design of the Control Unit
In the Von Neumann machine, data and instructions are transferred via the same data bus with a
single address bus for the addresses of the data and instructions, with the same word length in the
memory whether it's for data and instructions
Harvard Architecture:
The Harvard architecture typically separates data from instructions
- They are specially used in DSP’s – Digital Signal Processing – a type of embedded system
which specialises in audio/speech signal processing, sonar/radar signal processing, seismic
data processing etc
- Instructions may be held in read-only memory while the data may be stored in read-write
memory
- There could be a larger word size for instruction memory as some systems have much more
instructions than data
- Address bus may be wider than the data bus
- An embedded system is a special-purpose computer which is built into a device for a specific
function often operating in real time
- Because the Harvard architecture allows the instruction memory and the data memory to be
separate, both of these can be fetched parallel to each other which increases the speed of
the system
Von Neumann architecture:
Specifies the basic components of the computer and processor in which a shared memory and bus is
used for data and instructions
- The stored program concept can be defined as follows, machine code and instructions are
fetched and executed serially by a processor that performs arithmetic and logical operations
- The instructions are fetched sequentially and stored in the main memory, the sequence can
only change by a jump instruction
- For this to happen, the program must be within the main memory
- The instructions are fetched, decoded and executed by the processor one at a time
- This means one bus is for transferring data and instructions, this can lead to collisions and
bottlenecks in the computer. This is done as it simplifies the design of the Control Unit
In the Von Neumann machine, data and instructions are transferred via the same data bus with a
single address bus for the addresses of the data and instructions, with the same word length in the
memory whether it's for data and instructions
Harvard Architecture:
The Harvard architecture typically separates data from instructions
- They are specially used in DSP’s – Digital Signal Processing – a type of embedded system
which specialises in audio/speech signal processing, sonar/radar signal processing, seismic
data processing etc
- Instructions may be held in read-only memory while the data may be stored in read-write
memory
- There could be a larger word size for instruction memory as some systems have much more
instructions than data
- Address bus may be wider than the data bus
- An embedded system is a special-purpose computer which is built into a device for a specific
function often operating in real time
- Because the Harvard architecture allows the instruction memory and the data memory to be
separate, both of these can be fetched parallel to each other which increases the speed of
the system
