Inside the CPU: Fetch-Decode-Execute, Cores, Cache and Clock Speed

Hello students, welcome to Kwickprep. A laptop box says eight cores, three point two gigahertz and twelve megabytes of cache. What do those words really mean? Today we will see how a CPU runs a program, one tiny step at a time, and what makes one CPU faster than another.

Almost every computer today follows one design. It was described by John von Neumann in nineteen forty five. Its key idea is the stored program concept, which means the program is stored in memory, just like data. Data and instructions share the same memory and the same buses. The CPU then fetches and runs the instructions one after another.

Why was this such a big idea? Before it, early machines had to be rewired by hand for each new job, which took days. After it, you simply load a new program into memory. That means one machine can do many different jobs, just like your phone runs many apps.

This design has one weakness, called the von Neumann bottleneck. A bottleneck is a narrow point that slows everything, like the neck of a bottle. The CPU is far faster than main memory. Data and instructions travel on the same bus, one at a time. So the CPU often sits waiting for data, and cache memory was invented to help with this.

Before the cycle, meet the registers it uses. The program counter holds the address of the next instruction. The memory address register holds the address the CPU wants to read or write. The memory data register holds the data or instruction just moved from memory. The current instruction register holds the instruction being decoded. The accumulator holds the result of the latest calculation.

Now watch the cycle, one step at a time. First, the address in the program counter is copied to the memory address register. Next, the instruction at that address is fetched from memory into the memory data register. Then it is copied to the current instruction register, and the program counter goes up by one. The control unit decodes the instruction to find out what it means. Then the instruction is executed, for example the arithmetic logic unit adds two numbers. The cycle then repeats, millions of times every second.

Pause and predict. The program counter starts at one hundred. During the fetch, one hundred is copied to the memory address register, and the program counter becomes one hundred and one. What will the program counter and the memory address register hold at the next fetch? The memory address register gets one hundred and one, and the program counter becomes one hundred and two. The exception is a jump instruction, which loads a new address into the program counter.

Now let us see what makes a CPU faster. A core is a complete processing unit, and one chip can have many cores. More cores can run more tasks at the same time, like several billing counters in a supermarket. But it only helps when the software can split its work into parts. A quad core chip has four cores, and an octa core chip has eight.

Cache memory is very fast memory placed on or very close to the CPU. It keeps the data and instructions the CPU has used recently or uses often. There are levels called level one, level two and level three, and the smallest level is the fastest. A bigger cache means the CPU waits less for main memory, but cache is expensive. Think of it as the pen on your desk, instead of the one in your cupboard.

Inside the CPU, a clock sends out regular electrical pulses. Each pulse lets the CPU carry out a step of the fetch decode execute cycle. Clock speed is measured in hertz, and one gigahertz means one billion pulses every second. So a three point two gigahertz CPU gets three point two billion pulses a second. A higher clock speed also produces more heat, which is why CPUs need fans.

Let us compare the three factors side by side. More cores mean more tasks at once, but software must be able to split its work. More cache means less waiting for memory, but it is expensive. Higher clock speed means more steps every second, but it creates more heat. In exams, always mention both the benefit and the limit.

Finally, instruction sets. An instruction set is the complete list of basic commands that a particular CPU understands. Examples are add, load, store and jump. Each command is stored as a binary number called an operation code, or opcode. A program in machine code only runs on a CPU with a matching instruction set.

CPU makers follow two main styles of instruction set. Complex instruction set computers, called sisk, have many powerful commands, and they are found in most desktops and laptops. Reduced instruction set computers, called risk, use fewer and simpler commands that save battery. Almost every smartphone uses a risk design, and some laptops do too.

Let us revise what we learned today. In von Neumann architecture, programs and data share the same memory. The CPU fetches, decodes and executes each instruction, then repeats. More cores run more tasks, and more cache means less waiting. Clock speed is the number of pulses per second, measured in gigahertz. And an instruction set is the list of commands a CPU understands. Try tracing the cycle yourself on paper with the program counter at two hundred.