As computers become more ubiquitous by the minute, taking over every aspect of modern living, developing a deeper understanding of how they function has become essential.
Those who study or work in the technology sector need to know the ins and outs of a CPU, the Central Processing Unit.
But even as a consumer looking to purchase a new PC, knowing the CPU and its capabilities is paramount. Interestingly, a CPU is at the heart of many oft-used gadgets like a washing machine and a thermostat.
In this in-depth article, I cover the essence of the CPU in all its glory, discussing its components, functionality, evolution, and latest trends.
I will also discuss purchase and maintenance considerations from an end-user perspective.
The Central Processing Unit: What & Where is it?
As the name suggests, a CPU is the core unit of a computer, responsible for all the computational work as the ‘brains’ of the system. The CPU gets the instructions from memory and processes these signals. Then, it sends the output of the instructions to memory. It performs all the vital operations for running the operating system and the required applications, like a web browser or image editing software.
The CPU is usually in a socket on the motherboard, which is the primary circuit board of the computer.
The performance of the CPU gets measured in the number of instructions per second, also called IPS. A higher IPS indicates better speeds of task completion. Several factors determine this metric: the architecture, clock speed, and cache.
Why do you need a powerful CPU?
The power and capability of the CPU directly impacts the system performance. Consequently, a more powerful processor will allow your applications to launch faster and programs to run more smoothly. It is directly responsible for more enjoyable video gaming and other multimedia experiences.
Another key advantage of having a capable processor is staying updated with the latest software updates. An outdated CPU may fail to keep up, resulting in slower system speeds or freezes. Case in point: Only Intel Core processors eight-generation and higher can support Windows 11.
What is a CPU made of?
The CPU is essentially a silicon chip consisting of microscopic transistors. The chip is a tiny wafer of semiconducting material with an electronic circuit. CPU production uses photolithography, a technique where the image of the CPU gets etched onto the silicon chip. Inside are millions of transistors. The approach is called the ‘process node.’
The transistors act as switches, turning on or off based on the instruction that needs execution. Every computing task involves binary instructions — a set of zeroes and ones — and the transistors denote them. While ‘0’ denotes the off state, ‘1’ refers to the ‘on’ condition.
Here are the components making up the CPU:
1. Control Unit: The heart of the CPU
It connects the CPU to other computer components and handles the data flow between them. It has a decoder to interpret the instruction fetched from memory. The control unit changes the instruction into micro-operations, breaking it into executable tasks.
2. Registers for storage
They are the storage locations of the CPU where information about the current instruction gets stored. Registers are much faster than RAM or cache. CPUs have different types of registers, like general, instruction, program counter, accumulator, and index.
3. Arithmetic Logic Unit (ALU): The computation workhorse
The ALU conducts different operations — mathematical, like addition, subtraction, etc., and logical, like AND, OR, and NOT. As the worker unit, it gets data from the registers and processes it as directed by the control unit.
4. Clock
The CPU has a clock that generates a pulse at a stated frequency. It helps to synchronize all the operations, with the clock speed deciding how many instructions get executed per second. The measurement of the clock speed is in hertz (Hz).
5. Memory Management or Bus Interface Unit
This unit is in-charge of the interaction between the CPU and the memory. It also monitors the cache and virtual memory. The latter refers to space on the system’s storage (HDD or SSD) that gets treated as virtual RAM. It can help speed up currently running programs when the physical RAM falls short.
Is Cache Memory part of a CPU?
The cache is intrinsic to the performance of a CPU. It is much faster than registers and can fetch more instructions per second. The cache resides between the RAM and the CPU. It can store data that the processor requires from frequently visited addresses. The three main types are L1, L2, and L3 Cache.
Note that cache memory is much more expensive than registers. It is why the size of the cache plays a part in deciding the overall mettle of a CPU.
Understanding How a CPU Works: Step by Step
The functioning of a CPU happens in cycles. Every cycle has three steps.
1. Fetch the instruction
The CPU gets the instruction from the RAM or random access memory. It is in binary (0-1) form.
2. Decode
The instruction gets decoded. The CPU interprets the instruction and finds which component will be required to perform the task.
3. Execute
Finally, the instruction undergoes execution, and the output goes to the memory. It could be anything — from as elementary as playing a video to as complex as a mathematical function on a spreadsheet. The program counter gets updated to the address of the next instruction that the CPU will fetch.
This process is aptly called the fetch-execute cycle. It can take place millions of times every second.
Although a CPU processes instructions sequentially, modern technologies accommodate pipelining and parallelism to enhance performance.
What happens if the process gets interrupted?
As the CPU works on executing the instruction, it may encounter an interruption. It can be a signal from an external device, a task that demands urgent attention, etc. In this case, the CPU pauses the process and saves the state. It then executes the interrupted service routine. Finally, the original task is resumed.
Instruction Set Architecture (ISA): Defining how the CPU interacts with the software
The ISA is integral to the CPU, deciding how the hardware and the software interact. It is part of the abstract model and contains the instructions the processor can execute. New-age CPUs often have an extended ISA for tasks like virtualization, graphics rendering, and scientific simulation or modeling.
The Evolution of CPU Technology: Smaller is Faster
Increasing computing speeds has long been the cornerstone of technological evolution. But other factors have also guided the growth of processing technology.
- Reducing the size of the processing unit
- Limiting the power consumed
- Minimizing the amount of heat generated
The oldest CPUs were Vacuum Tubes
The first digital computer, ENIAC, was completed in 1945. It used vacuum tubes for processing — a whopping 18,000 of them. It also had thousands of resistors, capacitors, and joints, increasing the weight to over 30 short tons or 60,000 pounds. Not only was it unwieldy, but it also consumed 150 KW of electricity. The UNIVAC (Universal Automatic Computer), which came out in 1951, was also a vacuum-tube device.
Transistors and Integrated Circuits: 1940s to 60s
The invention of transistors in 1947 was the prime factor affecting a reduction in the size of the computer. The energy efficiency also improved manifold. In 1958, Jack Kilby of Texas Instruments patented the integrated circuit or IC. It combined many transistors onto one computer chip.
Microprocessors
The first microprocessor — the Intel 4004 — was invented by Intel in 1971. It could combine the whole CPU into one IC chip. The difference in the size of the computer was so incredible that personal devices became a reality.
From 16-bit to 64-bit microprocessors
The oldest microprocessors could work with 16-32 bits of data or 2-4 bytes. A 16-bit microprocessor has 16 address lines. Similarly, a 32-bit processor has 32-bit registers.
By 2001, Intel had produced a 64-bit model (with 64-bit registers). Unsurprisingly, these CPUs could handle much more complicated data processing. Today, most computers use 64-bit processors for modern operating systems like Windows 11 and Windows 10. However, they can work with 32-bit operating systems also.
Smaller but More Numerous Transistors
For modern central processing units, reducing the size of the transistors leads to an increase in the speed. With time, the size of transistors in CPUs has decreased, and their number has risen. Production trends indicate that the number doubles every two years — a generalization called Moore’s Law.
What is Moore’s Law?
This ‘law’ is a projection first stated by Mr Gordon Moore, the co-founder of Intel and Fairchild Semiconductor, in 1965. He predicted that the components in an integrated circuit would double each year. He revised it to every two years in 1975. Since then, this statement has been integral not only to CPUs but also to digital electronics. This exponential growth relates to better MOS technology (metal–oxide–semiconductor) — a type of field-effect transistor. It is also due to overall advancement in working with circuits and devices.
Single-core to Multi-core processors
After a point, increasing transistor density reaches a saturation level, making a further performance boost tricky. Manufacturers have moved from single- to multi-core CPUs, with multiple cores on one chip. When your CPU has numerous cores, it is equivalent to having many CPUs on a single chip.
Latest trends in CPU technology
1. Power Efficiency
The focus of modern-day processors is higher power efficiency. It has become imperative for devices like smartphones and portable laptops. Here are some ways this gets accomplished besides reducing the transistor size:
2. Advancing architectural design
It plays a significant part in improving energy efficiency. For instance, contemporary CPUs have smoother pipelining and can performparallel processing.
3. Better materials for semiconductors and integrated circuits
Silicon remains the most popular material for CPUs. However, manufacturers have started experimenting with newer materials like Gallium Arsenide (GaAs) and graphene. The latter is especially promising because of its high thermal conductivity and mechanical strength.
4. Newer power management techniques
AMD has come up with the CoolCore technology for more efficient performance. It activates or turns off components of the processor based on usage. Techniques like dynamic voltage scaling and power gating are also instrumental in managing power better.
5. The option of performance modes that consume less power
The TDP — Thermal Design Power — is a good measure of the performance of a CPU balanced with its efficiency. It denotes the heat generated, where a lower TDP rating is more efficient but can compromise performance. The TDP must be analyzed with other factors like the clock speed and the workload.
6. Turbo Boost and Overclocking
The focus of this technology, conceptualized by Intel, is improving CPU speeds without crossing limits of temperature and power. The concept is to allow the CPU to run at its base clock speed for regular workload. The speed adjusts dynamically during more intensive tasks. The result of this approach is an uptick in single-threaded and multithreaded applications. So, the Intel Core i9-13900K has a Turbo Boost frequency of 5.8 GHz even though the base is just 3 GHz.
A related technology is overclocking, which makes the CPU run faster than intended. Here, the processor’s clock rate rises, but not with dynamic adjustment like a Turbo Boost. Effectively, the boosted clock speed can become the base metric. While it delivers a performance boost, too much overclocking is detrimental. The absence of adequate temperature regulation can lead to burnout.
7. Accelerators
As artificial intelligence takes over more domains, CPUs have started using accelerators to handle these workloads. An AI accelerator is a parallel computation machine that makes AI-related processing, like neural networks, faster. It is an integrated circuit chip with billions of MOSFET transistors that cater to applications like robotics and the Internet of Things.
Intel is working on an advanced CPU that can run a generative AI chatbot with nil cloud dependence. It will make it more accessible to the end user to run language models or machine learning applications.
8. Quantum computing
This emerging technology focuses on quantum mechanics to deal with problems classical computers struggle to handle. Quantum processors use ‘qubits’ and not standard bits to execute multidimensional algorithms. They can assist with cybersecurity, material science, and machine learning challenges.
Platforms like Amazon Web Services have options for hosting, letting you pick customized resources, including the CPU, RAM, and storage. The cloud provider supports Intel, AMD, and Arm processors.
Performance Modes and How they Impact CPU Speeds
Many leading PC manufacturers have now introduced customized performance modes for your computer. These modes work per the current need of the hour, upping performance during gaming and slowing things down when conserving battery is paramount. For example, in Dell gaming PCs, performance mode ups processor speeds and cooling rate.
ASUS has pioneered an Intelligent Performance Technology (AIPT) focused on increasing CPU performance without compromising battery life, generating too much noise, or producing excessive heat. It runs specialized algorithms and sensors coupled with aerodynamic fan blades to produce these results. In contrast, its Balanced and Whisper modes are for everyday and quiet, power-saving tasks respectively.
CPU Architecture: Main Types
Computers can use different CPU architectures, i.e., how the CPU is designed and organized. The two main types are below:
1. CISC or Complex Instruction Set Computer
This architecture results in high-performance processors for personal and day-to-day use. One instruction can perform many actions, taking multiple clock cycles to complete. Intel’s x86 is a popular form of CISC architecture.
2. RISC or Reduced Instruction Set Computer
Here, the instruction set is simple, which consumes lower power and generates faster speeds. This architecture is also adept at multitasking. The idea is to execute one instruction per clock cycle. RISC is preferred for servers, embedded systems, etc. Arm and TriCore are some examples of RISC, the former having been a focal point for Apple’s research in recent times. The M1 chips for some new MacBooks are Arm-based.
RISC-V is also becoming popular lately, offering customization and application-specific extensions. It is similar to RISC but is open, allowing developers to contribute freely. It brings more flexibility to the system.
CPU Cooling Techniques
Considering all the hefty work that a CPU performs, keeping it cool is mandatory. Heat gets generated during processing and can build up if not dissipated, leading to a performance drop and lowered sustainability.
The conventional approach is to distribute the generated heat to the lid of the CPU and then to the baseplate of the cooler. Finally, a liquid or a heat pipe wards the heat to a fan, diverting it from your computer. Two methods are standard for today’s processors:
- Air Cooling relies on heat pipes to transfer the heat to a sink. The heatsink has metal fins that ward off the heat with cooler air and an attached fan. The effectiveness of this approach depends on the material of the pipe — copper or aluminum — and the number of fans used.
- Liquid cooling uses a coolant to absorb the heat. It travels to a radiator, from where fans dispel it. The standard coolant is purified or deionized water.
Broadly speaking, air cooling is cheaper and easier to install. However, liquid cooling is quieter and more efficient. The method you opt for will depend on your budget and usage patterns.
Do all computing devices have CPUs? What is an SoC?
Most modern-day PCs and laptops have central processing units. However, some tablets have an SoC or a ‘System on a Chip.’ An SoC contains the CPU and also components like graphics chips and memory. So, an SoC can perform more than computing functions typical to a standard CPU. It is compact and lightweight and does not need much power. You can also find SoCs in electronic gaming devices, smartwatches, robotic devices, and automobile navigation.
Buying a CPU: What all should you consider?
Inspecting the details of the CPU is pivotal when selecting a new desktop or laptop. It will determine the computing speed and efficiency you can expect during your day-to-day tasks like business operations, editing A/V content, gaming, or wireless streaming.
1. Number of Cores
Originally, CPUs were designed with one core. But now, they often consist of multiple cores, from 2 to over 32. For example, the Intel Core i9-13900KS processor has 24 cores. The AMD Ryzen 7 5800X from the 5000 Series has 8. Generally, computers designed for gaming an multimedia content editing need more cores for faster computation and multitasking.
2. Number of Threads
Multithreading is a popular technology for many contemporary CPUs. It means multiple threads can run in a single process. So, if a CPU supports two threads, one physical core can run two lines of execution at once.
Hyper-threading has also emerged recently, wherein one physical processor breaks into two virtual processors. It is a form of simultaneous multithreading.
How are Threads and Cores Different?
Threads are virtual cores. They are a logical software unit that run within a core. So, if a CPU can execute more threads, it completes tasks faster.
Since threads are virtual, the difference they bring to overall performance isn’t comparable to physical cores. However, multithreading is popular because of the upswing it lends to efficiency.
Multithreaded processors are especially helpful for gamers. Titles like ‘The Last of Us’ and PUBG utilize several threads. Typically, gaming enthusiasts will do well to get CPUs with 6-8 threads. Going for a multi-core processor with hyper-threading seems like the optimal solution, but you should weigh the cost, power consumed, and the benefits you can reap from the arrangement. Some games like Minecraft emphasize single-threaded performance.
3. Clock Speed
It is a critical metric for deciding the performance of the CPU. Many manufacturers highlight it when you look for a new one. The clock speed is the number of instructions the unit can execute in one second. Gigahertz (GHz) is the measurement unit.
I advise caution when using the clock speed to decide which CPU is faster. If you purchase an older CPU (from a previous generation) with a higher GHz number, the resulting speed may still be slower than a new-generation CPU with a lower clock speed figure. The silicon used in the chip keeps advancing with production cycles and improves performance.
You can also find CPUs with variable clock speeds. The speed varies with the workload to balance the performance and the power consumption.
4. Lithography
What is a 7nm or 10nm CPU?
CPUs often get marketed based on their nm measurement, which stands for nanometer. For example, a 7nm CPU means the size of the transistors is 7 nanometers. Intel’s Meteor Lake architecture is based on 7nm manufacturing. AMD Ryzen has announced a 7000 series of processors based on 5nm architecture.
The transistor density — 7nm, 10nm, or 5nm — plays a part in the processor speeds. Effectively, 7nm is more advanced and efficient than 10nm, improving the performance due to the higher transistor density. CPUs built via an advanced process node are likelier to maintain lower temperatures during intensive tasks. It helps the CPU attain boosted clock speeds for longer than a processor with an older process node. The TDP is also more manageable since less power is required.
However, please exercise caution since lithography is not the sole factor determining CPU performance. It is salient to separate fact from marketing gimmicks when seeking a competent processor.
5. Cost: The latest generation costs more
The money you spend on the CPU is dependent on your usage. Those who run resource-intensive applications, like professional gamers and creative artists, will need a more powerful CPU with more cores, preferably from the latest generation. Regular users whose routine spans creating documents, checking e-mail, or streaming multimedia content can get by with a cheaper and less powerful CPU. The prices range from around $200 to over $600. Assessing the price-to-performance ratio will help you make a considered purchase decision.
Intel’s latest line of CPUs is from the 13th generation. The i9-13900K is among the best processors you can buy. It is designed on the Raptor Cove microarchitecture and has both performance and efficiency cores. If your needs are not intense, you can opt for an older model, like an Intel Core i5 from the 12th generation. Likewise, AMD has released the AMD Ryzen 9 7950X — a highly competent option that falls behind its Intel counterpart by a thin margin.
6. Brand Name
The biggest names are Intel and AMD, with the former more established for its leading performance, efficiency, and single-thread clock speeds. However, AMD processors can be cheaper and are very close in most performance metrics. They can also deliver advantages in battery backup and multitasking.
Among the leading CPU manufacturers for smartphones and tablets is Qualcomm. While NVIDIA is known for its graphics cards, now infused with technologies like ray tracing, the company also makes CPUs for artificial intelligence.
CPU or GPU: What to prioritize?
A key consideration while purchasing a new laptop is whether to look for a more powerful CPU or GPU (in case you don’t have the budget for both). The GPU, or the graphics processing unit, is a specialized electronic circuit that came out in the 1990s. Notably, it is not the same as a graphics card. Instead, a GPU is the primary component of a graphics card. The card stays in a separate board on the motherboard. It is not built into the computer, making it interchangeable.
Initially, GPUs attended to faster image processing. But they have also been found to be useful for non-graphics tasks like machine learning and mathematical/scientific calculations. A GPU is built in a way that lets it conduct an operation on multiple data points in parallel.
Discrete graphics
Opting for a premium discrete graphics card from the NVIDIA RTX 40 Series is perfect for users whose needs revolve around high-end gaming or AI-based computation. The AMD Radeon RX series also has splendid picks.
Integrated CPU and GPU
Selecting integrated graphics is a good choice for cost-effectiveness. Here, graphics are built onto the same chip as the CPU. This combination also delivers energy efficiency compared to dedicated graphics processors. Intel Iris Xe graphics or AMD Radeon are the most favored choices.
Hybrid architecture
Intel has designed Core Ultra processors with inbuilt GPU and NPU (neural processing unit). The purpose is to optimize gaming and streaming while also maximizing AI acceleration.
Maintaining Your CPU
No matter how capable your CPU is, it can fall prey to conditions like overheating. Chances are something is amiss with the unit if you experience a performance drop, all other things remaining constant. The typical issues that can occur with a CPU are:
- Aging
- Excessive heat generation
- Inadequate power
- Dirt accumulation in the fans
- Malware
- Insufficient ventilation
I recommend maintaining the central processing unit by keeping your system clean and well-ventilated. You should also use an anti-virus application to protect against security threats. A straightforward way to optimize CPU performance is by cutting down on applications you don’t need.
Your possibilities expand if you have built your system and know how to work with hardware. For instance, you could replace the thermal paste responsible for cooling. You can use an isopropyl wipe to clean it.
The Computing Ecosystem: The CPU is a part of the whole
While the central processing unit is at the heart of your computer, it cannot stand alone. You will not find much benefit from a high-speed CPU if your system does not have enough RAM and storage. For instance, many gamers who buy a recent Intel Core i9 processor (like the Intel Core i9-14700K) will also seek at least 32GB of RAM and a 1TB or more SSD.
Generally, 16GB is the recommended RAM to support most modern games and other daily tasks; it is also reasonably affordable. Some enthusiasts can acquire as much as 64GB of RAM, while 4GB lies at the extreme skimping end.
Similarly, you should consider the storage options carefully to maximize the CPU performance. SSDs, or solid-state drives, are all the rage now, noted for their faster read speeds and overall snappiness. But you can also opt for a hybrid drive that combines an HDD (traditional hard disk drive) and an SSD.
Final Word
Newer components like GPU and NPU have assumed rising importance in the modern world. However, the CPU remains critical; the computer won’t work without it as the BIOS, or the basic input/output system, needs it to function. The CPU delivers power to all the components, acting as the brain of your system. Selecting the appropriate CPU for your computing needs and understanding its core functions is pivotal to deriving the optimal benefit from your electronic device.