3.2 Key Computer Components
Learning Objectives
After studying this section, you should be able to:
- Identify the major components of a modern computing system.
- Explain the roles of the CPU, GPU, RAM, and storage.
- Distinguish between volatile and nonvolatile memory.
- Interpret common measures of storage capacity and data transmission.
- Explain Moore’s Law as a historical trend and discuss why computer performance depends on more than processor speed.
- Consider how improvements in computing affect business decisions.
What Is Inside a Computer?
Computers come in many forms. A smartphone looks different from a laptop, a gaming system, or a server in a data center. But most computing devices rely on the same basic functions: process information, temporarily hold information, permanently store information, and communicate with other devices. A useful way to think about a computer is as a system of cooperating components rather than as one powerful “brain.”
The Basic System
Input → Processing ↔ Memory → Storage → Output / Network
For example, when you open an Excel workbook:
- The file is retrieved from storage.
- The needed data and program instructions are placed in RAM.
- The processor performs calculations and executes Excel’s instructions.
- Results appear on the screen and may be saved back to storage or transmitted through a network.
No single component does all of this by itself. Let’s see how the components all work together!
The von Neumann architecture, one of the foundational models for how a computer is organized. The architecture centers on two major parts: the Central Processing Unit (CPU) and the Memory Unit. The CPU contains the Control Unit and the Arithmetic/Logic Unit (ALU). The Control Unit directs the computer’s operations by fetching and interpreting instructions, while the ALU performs calculations and logical comparisons.
The Memory Unit stores both data and program instructions. The two-way arrows between memory and the CPU show that information moves back and forth: the CPU retrieves instructions and data from memory, processes them, and can send results back to memory. The key idea of von Neumann architecture is that instructions and data are stored in the same memory system. This stored-program concept allows one computer to perform many different tasks simply by loading different programs.
The CPU fetches an instruction from memory, the Control Unit decodes it, the ALU or another CPU component carries it out, and the result may be stored back in memory. This diagram is simplified because a full von Neumann model would usually also show input and output devices and the pathways, or buses, that connect the CPU, memory, and other components.
CPU: The General-Purpose Processor
The central processing unit (CPU) executes instructions and coordinates much of the work performed by a computer. The CPU performs operations such as:
- arithmetic calculations,
- comparisons and logical decisions,
- moving data,
- executing program instructions,
- and coordinating other hardware.
For many years, processor performance was commonly described primarily by clock speed, measured in gigahertz (GHz). Clock speed still matters, but it is no longer a good way to compare computers by itself. Modern processors contain multiple cores. A core is an individual processing unit capable of executing instructions. A processor with several cores can perform multiple tasks at the same time when its software is designed to take advantage of them. A computer with a lower clock speed but more efficient architecture, more cores, better memory, or specialized processors may outperform a computer with a higher advertised GHz number.
When comparing two computers, avoid assuming that higher GHz means a faster computer. A computer’s performance depends on the entire system and on what you are asking the computer to do. A computer used primarily for spreadsheets and web applications has different requirements from one used for video editing, gaming, engineering simulations, or training artificial intelligence models.
GPU: Built for Parallel Work
A graphics processing unit (GPU) was originally designed to perform the many calculations required to generate graphics. Graphics involve performing similar calculations on enormous numbers of pixels, shapes, textures, and objects at the same time. GPUs therefore developed differently from CPUs. Instead of emphasizing a smaller number of powerful general-purpose processing cores, GPUs contain many processing units designed to perform large amounts of work in parallel. That design turned out to be useful for much more than graphics.
GPUs are now widely used for:
- gaming and graphics,
- video production,
- scientific computing,
- simulations,
- data analytics,
- machine learning,
- and generative AI.
This does not mean that a GPU replaces a CPU. Modern systems increasingly use the CPU and GPU together, assigning different types of work to the processor best suited to the task. You will see how this shift changed the direction of computing in Section 3.3.
AI Accelerators and Neural processing units (nPUs)
Many newer devices also contain processors designed specifically for artificial intelligence workloads. A neural processing unit (NPU) or other AI accelerator performs mathematical operations commonly used by machine-learning systems more efficiently than a general-purpose CPU.
These processors may handle tasks such as:
- image enhancement,
- speech recognition,
- background removal during video calls,
- language processing,
- and locally running AI features.
This is part of a larger trend called heterogeneous computing: instead of asking one type of processor to do everything, a system can contain different processors designed for different jobs. OpenStax discusses this shift toward CPUs, GPUs, TPUs, and other specialized processor architectures in Section 5.6, Processor Architectures.
RAM: The Computer’s Working Space
Random-access memory (RAM)
provides fast temporary storage for programs and data currently being used. When you launch an application, information needed by that application is generally moved from longer-term storage into RAM because the processor can access RAM much more quickly. Think about working at a desk. Your filing cabinet may hold hundreds of folders, but you pull the documents you are actively using onto the desktop. A larger desktop allows you to keep more material immediately available.
More RAM can help a computer:
- keep more programs open,
- work with larger files,
- switch among applications,
- and avoid repeatedly retrieving information from slower storage.
RAM is normally volatile memory, which means its contents disappear when the device loses power. That is why an unsaved file may disappear after a computer crashes or shuts down unexpectedly.
Storage: Keeping Data After the Power Is Off
Long-term storage is nonvolatile, meaning that it retains information when power is removed. Most personal computers today use a solid-state drive (SSD). SSDs store information electronically and contain no mechanical moving parts. Compared with traditional hard disk drives, SSDs generally provide faster access, use less power, operate silently, and are more resistant to physical shock.
Traditional hard disk drives (HDDs) store information magnetically on spinning disks. They remain useful when very large amounts of storage are needed at relatively low cost. Other forms of nonvolatile storage include:
- USB flash drives,
- memory cards,
- external SSDs and hard drives,
- and storage provided through cloud services.
Bits, Bytes, and Data Capacity
Computers represent information digitally using bits. A bit represents a binary value: 0 or 1. Eight bits make one byte. Storage capacity is normally expressed in bytes, while network transmission speeds are normally expressed in bits per second.
| Unit | Approximate Size | Common Example |
|---|---|---|
| 1 byte (B) | 8 bits | One character of simple text, such as A |
| 1 kilobyte (KB) | 1 thousand bytes | A short plain-text document or a few paragraphs |
| 1 megabyte (MB) | 1 million bytes | A compressed photo or about a minute of compressed audio |
| 1 gigabyte (GB) | 1 billion bytes | A few hundred smartphone photos or roughly 1–2 hours of compressed HD video |
| 1 terabyte (TB) | 1 trillion bytes | About 250,000 photos or hundreds of hours of HD video |
| 1 petabyte (PB) | 1 quadrillion bytes | About 1,000 terabytes; enough for very large organizational data collections |
| 1 exabyte (EB) | 1 quintillion bytes | About 1 million terabytes; useful for describing global-scale data volumes |
| 1 zettabyte (ZB) | 1 sextillion bytes | About 1 billion terabytes; used to describe data generated or stored at worldwide scale |
For everyday purposes, the approximate decimal definitions above are usually sufficient.
Storage vs. Speed
Suppose your internet service is advertised as 1 Gbps. That means approximately 1 gigabit per second, not one gigabyte per second. Because there are eight bits in a byte: 8 bits = 1 byte. A network measurement written with a lowercase b usually refers to bits. A storage measurement written with an uppercase B usually refers to bytes. That tiny change in capitalization matters.
Input, Output, and Networking
A useful computer also needs ways to interact with people and other systems. Input devices send information into a system. Examples include:
- keyboards,
- touchscreens,
- cameras,
- microphones,
- scanners,
- and sensors.
Output devices
communicate information from the system. Examples include:
- displays,
- speakers,
- printers,
- and haptic feedback.
Increasingly, computers also communicate directly with other computers. Network interfaces allow devices to connect through Wi-Fi, cellular networks, Ethernet, Bluetooth, and other technologies. In modern business systems, networking may be just as important as the processor itself. A cloud-based application may perform much of its processing hundreds or thousands of miles from the person using it.
The importance of Moore’s Law
For decades, improvements in computing were strongly associated with Moore’s Law, named for Intel cofounder Gordon Moore. In 1965, Moore observed that manufacturers were able to place rapidly increasing numbers of electronic components on integrated circuits. The observation later became associated with the tendency for transistor counts to roughly double over a period of about two years. A transistor is a microscopic electronic switch. Modern processors contain billions of them.
Moore’s Law is therefore best understood as a historical observation about the increasing density of components on computer chips—not as a rule stating that every new computer becomes twice as fast on a fixed schedule. For many years, smaller and more numerous transistors contributed to impressive improvements in speed and cost. But processor performance now depends on many other factors as well. Modern computing gains increasingly come from:
- multiple processing cores,
- GPUs and other accelerators,
- improved processor design,
- faster memory and storage,
- better software,
- specialized AI processors,
- networking,
- and combining many computers together.
The next section traces how that transition occurred. For a contemporary discussion of processor architecture, multicore computing, GPUs, and specialized processors, see OpenStax, Introduction to Computer Science, Section 5.6.
Why These Improvements Matter to Business
The underlying technology may be technical, but its effects are not. Cheaper and more capable computing changes what businesses can afford to do. A capability that once required specialized equipment or a major technology company may eventually become available to a small business or individual employee. Examples include:
- high-quality digital photography,
- cloud data storage,
- videoconferencing,
- GPS navigation,
- data visualization,
- automated translation,
- predictive analytics,
- and generative AI.
Falling costs can create new products and markets while disrupting older ones. Managers therefore do not need to become semiconductor engineers, but they do need to recognize the direction of technological change. A business deciding whether to purchase hardware, move applications to the cloud, adopt AI systems, automate a process, or replace equipment is making decisions influenced by the economics of computing. Technology also depreciates quickly. The most expensive or powerful device available today is not automatically the best business investment.
Key Takeaways
- A modern computer is a system of cooperating components, not simply a CPU.
- The CPU handles general-purpose processing and coordination.
- Modern CPUs contain multiple cores.
- GPUs specialize in highly parallel computing and are important for graphics, analytics, and AI.
- NPUs and other accelerators perform specialized tasks efficiently.
- RAM is fast, temporary, volatile working memory.
- SSDs and other storage devices provide nonvolatile long-term storage.
- Storage is usually measured in bytes, while network speed is usually measured in bits per second.
- Moore’s Law describes the historical growth in component density on integrated circuits; it does not mean computer speed automatically doubles every eighteen months.
- Modern performance improvements increasingly come from combining different processors, better software, networking, and large computing systems.
References and Attribution
- Franchitti, Jean-Claude. Introduction to Computer Science. OpenStax, 2024.
Section 5.6: Processor Architectures. - Franchitti, Jean-Claude. Introduction to Computer Science. OpenStax, 2024.
Section 4.3: Parallel Programming Models. - Raisinghani, Mahesh S. Foundations of Information Systems. OpenStax, 2025.
OpenStax Foundations of Information Systems.
Section Attribution
This section is an original instructional synthesis written for introductory business information-systems students. It incorporates and adapts concepts from the openly licensed OpenStax sources linked above. OpenStax material is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0). Access for free at: OpenStax: Introduction to Computer Science
and OpenStax: Foundations of Information Systems.
