What is Processor: Components and Types

A processor is the most important integrated circuitry (IC) in computer. It is used for interpreting most computer commands. It is an electrical component that performs operation on external data sources such as memory and data stream. Typically taking the form of microprocessors, it can be implemented on integrated circuit chips.

A processor, also known as a central processing unit (CPU), is the primary component of a computer that performs most of the processing inside the computer. It acts as the “brain” of the computer, executing instructions from computer programs.

Table of Contents

• What is processor?
• Components of processor
• How do processors work?
• Types of processor

What is Processor in Computer?

Processor or Central Processing Unit (CPU) refers to logic circuitry that responds to and processes basic instructions that drive computers. This integrated electronic circuit performs calculations that run computers. These circuits are found in electronic devices. They receive input in the form of program instructions and execute calculations for providing with which the user will interact. 

For any operation on a computer, the processor must interpret the operating system. A processor consists of arithmetical logic and a control unit (CU) that measures capability in terms of:

• Maximum number of bits/instructions
• Relative clock speed
• The ability to process instruction at a given time

Components of Processor

A processor has four components: a floating point unit (FPU), an arithmetic logic unit (ALU), registers, and cache memories. 

1. Arithmetic Logic Unit (ALU)

ALU is the main component in a processor that performs various arithmetic and logic operations. It is an integrated circuit within the CPU/GPU, due to which it is also known as an integer unit (IU). This is the last component that performs calculations in the processor. 

2. Floating-Point Unit (FPU)

It is part of the computer system used for carrying out operations on floating-point numbers. These operations include square root, multiplication, division, subtraction, and addition. It can perform transcendental functions such as trigonometric and exponential functions; however, it may not be accurate. 

3. Registers

Registers are types of computer memory that accept, transfer, and store data and instructions being used. It instructs ALU about the processes that must be carried out and stores the results of these operations. 

4. Cache

Cache is the smaller yet faster memory located close to the processor’s core. This memory stores the copy of data from the frequently used main locations. There are three levels of cache: L1, L2 and L3 cache. L1 is the primary chip, which is embedded in the processor chip.

Since it is small, it has limited storage. L2 cache is the secondary cache that is either embedded on a processor chip or a separate chip with a high-speed bus that connects it to the CPU. Also known as processor cache, L3 is a specialized backup memory for L1 and L2. It boosts the performance of L1 and L2.  

How Do Processors Work?

Every processor is constituted of one or more individual processing units called 'cores'. Every core processes instructions from a single computing task at a particular speed which is also known as 'clock speed'. This speed is measured in gigahertz (GHz). In modern computers, there are multiple processor cores since increasing clock speed is a technically difficult task after a point. 

All these processor cores work together for processing instructions and completing multiple tasks at the same time. Within the processor, there are four key elements including arithmetic logic unit (ALU), floating point unit (FPU), registers and cache memories. ALU carries basic and advanced arithmetic and logic operations. These results are sent to registers that also store instructions. Cache are the small and fast memories that store data copies for frequent use.

To process instructions, processors use an instruction pipeline. This typically involves stages such as fetch (retrieving the instruction from memory), decode (determining what the instruction does), execute (performing the actual operation), and write-back (storing the result). To optimize execution, processors employ branch prediction techniques, attempting to guess the outcome of conditional instructions before they're fully processed.

Processors interact with a larger memory hierarchy beyond their internal components. This starts with multiple levels of cache (L1, L2, L3), then extends to main RAM, and even storage devices. Each level in this hierarchy offers a trade-off between speed and capacity, with faster but smaller memories closest to the processor.

In modern computers, multiple processor cores work together to handle various tasks simultaneously. This is complemented by multi-threading capabilities, allowing each core to handle multiple threads of execution at once, further improving performance and efficiency.

Different Types of Processors

Let us now discuss the different type of processors that are available at present.

1. Application-Specific Instruction Set Processor (ASIP)

It is a component used in the system-on-a-chip design. The instruction set of ASIP is customized to benefit specific applications. For certain ASIPs, this instruction set is configurable. ASIP can be an alternative to hardware accelerators for video coding or baseband signal processing.  

Explore networking and hardware courses

2. Processor based on Flynn’s classification

According to Flynn’s taxonomy, processors can be classified based on concurrent instructions and data streams available in architecture. Let us now discuss each of the above one by one.

2.1 SISD (Single Instruction Single Data) 

It is a computer architecture in which a single uni-core processor executes a single instruction stream. This is done for operating on data stored in a single memory. SISD can have (as per Flynn) concurrent processing characteristics. 

Here, instructions are sent to the control unit from the memory module. Then, they are decoded and sent to a processing unit that processes data retrieved from the memory module and then sends it back. Examples of SSID are traditional uniprocessor machines such as PCs, old mainframes, pipelined, and superscalar processors.

2.2 Single Instruction Multiple Data (SIMD)

It is a type of computer that comes with multiple processing elements. It simultaneously performs same operation on multiple data points as well as parallel computations on only a single instruction at a given time. SIMD may be a part of the hardware design and is directly accessible through instruction set architecture (ISA). These machines do not exploit concurrency.

2.3 Multiple instruction Single Data (MISD) 

It is a type of parallel computing architecture where multiple functional units perform different operations on the same data. Every CU here handles and processes one instruction stream through corresponding processing elements. It has an architecture that is used for fault tolerance. MISD organisation computers are used rarely. Space Shuttle flight control computer is an example of MISD.

2.4 Multiple Instruction Multiple Data (MIMD) 

This refers to a technique used for achieving parallelism. Machines that have MIMD have several processors that function independently and asynchronously. Multiple autonomous processors execute, at any time, execute different instructions on different data pieces. 

These machines can be either shared or distributed memory categories based on how MIMD processors access memory. Shared memory may be bus-based, hierarchical or extended types. Distributed memory may be of hypercube or mesh types.

3. Processor Based on Number of Cores

Following are the different types of processors based on the number of cores:

3.1 Single core

A single-core microprocessor has a single core in its die. It performs the ‘fetch-decode-execute cycle’ once per clock cycle since it runs only on a single thread. These processors have been less in demand due to lesser processing power. Their slow speed has made multi-core systems more popular. 

3.2 Multi-core

Multi-core processors are microprocessors on a single integrated unit having two or more cores. Each core reads and executes the program instructions. Here, a single processor can simultaneously run instructions on a separate core. Due to this, the overall speed for programs supporting multithreading and parallel computing techniques increases.

3.3 Hyper-Threading

It is a technology that is used in the Intel microprocessors. This technology allows a single microprocessor to act as two processors for the operating system and the application. Through hyperthreading, processor resources are more efficiently used, allowing multiple threads to run on each core. 

4. Special processors 

Following are the different types of special processors:

4.1 Graphics processing unit (GPU)

It is a specialized electronic circuit that manipulates and alters memory to accelerate the creation of images in frame buffers that are intended for output to display devices. They can efficiently manipulate image processing and computer graphics. 

Due to their highly parallel structure, they are more efficient than general-purpose CPUs for those algorithms where processing of large data blocks is performed in parallel. GPU may be embedded on motherboards or video cards.  

4.2 Physics Processing Unit (PPU) 

Also known as the Physics Acceleration card, it is a dedicated microprocessor that handles physics calculations, unlike GPU. It is used specifically for the physics engine of video games. This microprocessor helps offload time-consuming tasks for the computer’s Central Processing Unit. It provides physics simulation data and communicates this data to the CPU. These are used in high-performance computers. 

4.3 Digital Signal Processor (DSP) 

It is another specialized microprocessor having an architecture optimized for the operational needs of digital signal processing. This measures, compresses, and filters continuous real-world analog signals. They are more power efficient, due to which they can be used in portable electronic data. 

These processors fetch multiple instructions and data at the same time. DSPs are cost-effective since they are cheaper yet provide better performance and lower latency. They do not have any requirements for specialized cooling or larger batteries. 

4.4 Network processor

It is a special-purpose hardware device that is programmable. Like RISC processors, these are low-cost and flexible, scalable and fast as ASIC chips. Such processors are used for designing networking applications. 

They have characteristics that are similar to general-purpose CPUs used in different types of equipment and products. Firewalls, routers, switches and network security devices use network processors. 

Explore networking courses

4.5 Front end processor

These are smaller computers that connect networks to host computers. Data is transferred between the front-end processor and the host computer through high-speed parallel interfaces. 

They offload the host computer from managing peripheral devices, packet assembly and disassembly, and error detection and correction. These processors communicate with peripheral devices using serial interfaces via communication networks.