You write Python.
But Python is not executing your code the way you think.

Today, you stop treating Python as magic.

Today’s Goal

By the end of today, you will:

  • Understand how Python executes your code
  • Learn what happens between .py to execution
  • Build a mental model of the Python interpreter

The Illusion

You write:

x = 10
y = x + 5
print(y)

You assume:

Python reads this line by line and runs it

That assumption is wrong.

The Real Pipeline

Your code goes through three stages:

Source Code (.py)
|
v
Compilation -> Bytecode (.pyc)
|
v
Execution -> Python Virtual Machine (PVM)

Step 1 - Compilation

Before execution, Python converts your code into:

Bytecode

This is a low-level instruction set designed for the interpreter.

Example (Conceptual Bytecode)

LOAD_CONST 10
STORE_NAME x
LOAD_NAME x
LOAD_CONST 5
BINARY_ADD
STORE_NAME y
LOAD_NAME y
PRINT

This is what Python actually executes, not your original code.

Where Is This Stored?

Python may generate:

__pycache__/
    file.cpython-311.pyc

This .pyc file contains:

  • compiled bytecode
  • optimized for faster startup

Important Insight

Python is compiled to bytecode, then interpreted.

Step 2 - The Python Virtual Machine (PVM)

This is the engine that runs your code.

Execution Loop (Simplified)

while instructions exist:
    fetch instruction
    execute instruction

What This Means

Every line of your code becomes:

  • multiple instructions
  • executed one by one
  • inside a loop

Key Insight

Python execution is a loop over bytecode instructions.

Why This Matters

Because:

  • Each operation has overhead
  • Python is slower than compiled languages
  • Performance is tied to instruction count

Inspect It Yourself

Use the built-in disassembler:

import dis

def f():
    x = 10
    return x + 5

dis.dis(f)

Example Output

LOAD_CONST 10
STORE_FAST x
LOAD_FAST x
LOAD_CONST 5
BINARY_ADD
RETURN_VALUE

One line becomes multiple operations.

Mental Model Upgrade

Before: Python runs my code

Now: Python executes bytecode instructions in a loop

Hidden Cost

Even simple operations involve:

  • stack operations
  • reference handling
  • interpreter dispatch

Performance Insight

This is why:

  • Python loops are slower
  • function calls are expensive
  • optimized libraries use C

Your Task

  1. Write a small function
  2. Run:
import dis
  1. Observe:
  • number of instructions
  • how simple code expands

Common Mistakes

  • Thinking Python runs code directly
  • Ignoring bytecode layer
  • Assuming one line equals one operation
  • Not understanding execution cost

Think Deeper

  1. Why does Python compile to bytecode instead of machine code?
  2. Why is Python slower than C?
  3. What happens if .pyc file is deleted?

Final Insight

Python is a virtual machine executing instructions, not your source code.

Tomorrow

Memory model - objects, references, mutability

Rule

  • Build mental models
  • Question assumptions

See you in Day 2.