Why PyPy Runs Python Code Faster Than C: Understanding JIT Compilation

Python is praised for rapid prototyping, but its interpreted nature often makes it slower than compiled languages like C or C++. The article introduces PyPy as a solution that can make Python code run faster than C by using just‑in‑time (JIT) compilation.

A simple benchmark is presented: a loop that sums integers from 0 to 100,000,000 is executed with the standard Python interpreter, with PyPy, and with an equivalent C program. The results show CPython taking about 10 seconds, PyPy completing the task in 0.22 seconds, and the C version in 0.32 seconds, demonstrating that PyPy can even surpass native C performance.

The speed advantage stems from PyPy’s JIT compiler, which translates frequently executed Python bytecode into machine code at runtime, combining the flexibility of an interpreter with the performance of ahead‑of‑time (AOT) compilation.

In contrast, languages such as C, C++, Swift, Rust, and Haskell use AOT compilation, converting source code directly into machine code before execution. Interpreted languages like Python, JavaScript, and PHP keep source code unchanged and execute it line‑by‑line via an interpreter.

PyPy’s JIT approach compiles hot code paths into machine code before execution, offering the best of both worlds: the portability and ease of an interpreter with the speed of compiled code.

<code>import time
from termcolor import colored

start = time.time()
number = 0
for i in range(100000000):
    number += i
print(colored("FINISHED", "green"))
print(f"Ellapsed time: {time.time() - start} s")
</code>

Images in the original article illustrate the benchmark results and the concepts of AOT compilation, interpretation, and JIT compilation.