190. Cython for Performance

1

Basic Cython Speedup

• Convert a simple Python function to Cython for performance improvement.

example1.pyx

Copy

# Cython file (example1.pyx)
def add(int a, int b):
    return a + b

Compilation (Save as setup.py)

Copy

from setuptools import setup
from Cython.Build import cythonize

setup(
    ext_modules=cythonize("example1.pyx")
)

Run:

Copy

python setup.py build_ext --inplace

Use in Python:

Copy

from example1 import add
print(add(10, 20))  # Output: 30

---

2

Using cdef for Fast Loops

• cdef declares C variables for faster execution.

example2.pyx

Copy

cdef int square(int x):
    return x * x

def sum_of_squares(int n):
    cdef int i, total = 0
    for i in range(n):
        total += square(i)
    return total

Compile and use it in Python as shown in Example 1.

---

3

Using nogil for Parallelism

• nogil allows running code without Python’s Global Interpreter Lock (GIL).

example3.pyx

Copy

from cython.parallel import prange
cimport cython

@cython.boundscheck(False)
@cython.wraparound(False)
def parallel_sum(int n):
    cdef int i, total = 0
    with nogil:
        for i in prange(n, nogil=True):
            total += i
    return total

---

4

Using Cython with NumPy

• Speed up NumPy array operations with Cython.

example4.pyx

Copy

import numpy as np
cimport numpy as np

def sum_array(np.ndarray[np.float64_t, ndim=1] arr):
    cdef int i
    cdef double total = 0.0
    for i in range(arr.shape[0]):
        total += arr[i]
    return total

---

5

Calling External C Functions

• Use C functions inside Cython for extreme speedups.

example5.pyx

Copy

cdef extern from "math.h":
    double sqrt(double x)

def py_sqrt(double x):
    return sqrt(x)

Compile and call py_sqrt(25.0) from Python.

---