Basic Computer Theory

Learning assembly is not similar to learning yet another programming language. There are no high-level constructs. There is memory and instructions. That’s it.

Memory (RAM)

Imagine a room of personal-lockers inside a bank. Every locker has the same size, color, capacity, access mechanism and all the lockers are managed by an addressing system, which uniquely identifies them.

A locker can contain anything, but what identifies them commonly is valuables. A person can keep gold or silver items while another person can keep the photos of their family. Items are different but both of them identifies as a valuable.

Similarly, computer memory is a huge collection of boxes, where each box has the same capacity of 1 byte (or 8-bits) and all the boxes are uniquely identified by memory addresses.

Just as a locker itself can’t identify its contents, everything is just a valuable, the same is with memory.

• Everything is just raw bytes. What defines a byte as an integer, a decimal, an emoji, an alphabet is the interpretation of that byte (or a group of bytes).
• Previously, we have read that context and interpretation is what that rules assembly. We can see it in practice here.
• A byte can be interpreted as digit, as an alphabet. When those same bytes are grouped, and interpreted, their meaning changes.
• Context decides the kind of interpretation required in order to get the right meaning out of those bytes. That’s how the “illusion of data types” is created.

Note: The illusion of data types is quite dense and uncovering it is not as simple as it looks. Throughout this journey, we will understand it and uncover it. Remember not to hurry

Central Processing Unit

CPU is the physical component that actually executes instructions.

While modern systems are very complex, the role of CPU is quite simple: fetch, decode, and execute instructions stored in memory. And this what a fetch-execute cycle is.

The CPU needs to know the instruction to be executed, for this, it has got Program Counter, whose purpose is to hold the memory address of the next instruction to be executed.

The CPU has the instruction now and it needs to figure-out what this instruction means. For this, the CPU has instruction decoder.

In our everyday life, anything we do involves movement. We are moving things from one place to other. Locations are one the most important things in our life. And so as with CPU.

• Every instruction uses some memory locations, where some data is stored, which is required in the instruction. Instructions themselves can be stored at a memory location.
• To fetch that data from that memory location, what comes handy is the data bus. It is the connection between CPU and memory.

Suppose you have to go to your best friend’s house. You have picked up the keys for your bike, the cap, the scarf to protect from heat, and the sun glasses. All these things are stored somewhere in your house.

• The key is probably on the top of the fridge or in the key-holding area.
• The sun glasses are at the dressing table.
• The scarf is in the closet.
• And the cap is on the table.
• What about the address to your friend’s house? Where is that? In your mind?
• If you don’t go mad on me, can I ask why don’t you write the address of your friend on a paper and store it your closet?
• Obviously, you are saying, “I am not mad!” Keeping the address in your mind ensures that it is accessible all the time.
• The CPU also has some highly-efficient and rapidly-accessible locations for this exact purpose. They are called general purpose registers. These are high-speed memory locations inside the processor itself, that takes part in the actual execution.

At last, the Arithmetic and Logic Unit. The data and the decoded instruction is passed here for further processing. Here, the instruction is actually executed. The computed results are placed on the data bus and sent to the appropriate location (a memory, or a register), as specified by the instruction.

And these are the core elements of the CPU.

Fetch-Execute Cycle

1. Fetch – Read the instruction from memory (address held in instruction pointer).
2. Decode – Understand what the instruction means.
3. Execute – Perform the operation (move data, add, compare, etc).
4. Store – Write the result (often into a register or memory).
5. Repeat – Move to the next instruction.

And this happens billions of times.

Registers

In addition to the memory outside of the processor, the processor itself has some special, high-speed memory locations called registers.

The primary purpose of a register is to hold the data that the CPU is actively working on. Just like the desk has a pen-holder which holds necessary pens like black and blue ball pens, a correction pen, a ruler, pencil, and a rubber. But it doesn’t contain the whole stationary. We have cupboards for that.

Since registers are high-speed and are located within the processor itself, they are limited in number, for various valid reasons.

Most information is stored in the main memory, brought in the registers [,for processing], and then put back into memory when the processing is completed.

Just like memory, registers also hold bits. It’s up to you to interpret them correctly (as numbers, characters, addresses, etc.)

Mainly, there are two types of registers:

• General Purpose Registers, this is where the main action happens. Addition, subtraction, multiplication, comparisons, and other operations generally use GPRs. They are expansive and are very less in number. There are 16 GPRs in x86_64 (amd64 or 64-bit) architecture.
• Special Purpose Registers, self-explanatory.

Size/Width Of A Register

Word size is the size of the data, or the number of bits the CPU can process at once.

It is architecture dependent. 32-bit systems has a word size of 32-bits or 4-bytes, while 64-bit systems have 64-bit wide registers.

Word is the size of the registers in a particular CPU architecture.

“Word” is the fundamental unit of CPU. Just like m/s is for velocity.

Note: “Word” as a terminology is used quite loosely across different things. Therefore, remembering the context in which we are using it is paramount to prevent confusion. In this low level journey, the context is only assembly so anywhere we use “word”, refers to assembly only.

What is 32-bit && 64-bit?

We download software. If that’s a technical software, it is always mentioned whether you are at 32-bit or 64-bit OS. And there are different software for both.

Most modern systems are based on 64-bit architecture.

This 32-bit and 64-bit tells us how many numbers the CPU is capable of dealing at once.

• A 32-bit CPU can work with numbers that are 32 bits long.
• A 64-bit CPU can work with numbers that are 64-bits long.

ISA v/s Assemblers

There are multiple assemblers in the market. GNU assembler (GAS), Netwide assembler (NASM), fish assembler (FASM), Microsoft assembler (MASM) and so on….

Assembler does one thing — translate human readable assembly instructions into machine opcodes, which the CPU can understand and execute directly.

A CPU’s instruction set architecture defines its capabilities. It conceptualizes everything that the CPU can do.

Assemblers are the programs that decides how the programmers will interact with the CPU.

Take this, there are handful of firms that research on semiconductor chips, but there are relatively many who does the manufacturing. ISA is that research while assemblers are the manufacturers. Each manufacturer (assembler) has the freedom to manufacture its own way.

Assembly-Time vs Run-Time

Assembler directives mean nothing to the CPU. They exist to streamline development. The assembler resolves them into machine understandable things while assembling the source into object code. This is called assembly-time management.

There are things that are consistent across all the assemblers because the CPU directly understand them. These are runtime managed things.

For example, OFFSET is an assembler directive, while lea is a CPU understood operation, defined in the ISA itself.