How Numba and Cython speed up Python code

Last updated on February 10, 2018, in python

Over the past years, Numba and Cython have gained a lot of attention in the data science community. They both provide a way to speed up CPU intensive tasks, but in different ways. This article describes architectural differences between them.


Numba is a just-in-time (JIT) compiler that translates Python code to native machine instructions both for CPU and GPU. The code can be compiled at import time, runtime, or ahead of time.

It's extremely easy to start using Numba, by simply putting a jit decorator:

import numpy as np
from numba import jit

a = np.arange(1, 10 ** 7)
b = np.arange(-10 ** 7, -1)

def sum_sequence(a, b):
    result = np.zeros_like(a)
    for i in range(len(a)):
        result[i] = a[i] - b[i]
    return result
>>> fast_sum_sequence = jit(int64[:](int64[:], int64[:]), nopython=True)(sum_sequence)
>>> timeit.timeit('sum_sequence(a, b)', globals=globals(), number=1)
Basic Python version: 4.227093786990736
>>> timeit.timeit('fast_sum_sequence(a, b)', globals=globals(), number=1)
Numba version: 0.05048697197344154

As you may know, In Python, all code blocks are compiled down to bytecode:

>>> import dis
>>> dis.dis(sum_sequence)
  2           0 LOAD_GLOBAL              0 (np)
              2 LOAD_ATTR                1 (zeros_like)
              4 LOAD_FAST                0 (a)
              6 CALL_FUNCTION            1
              8 STORE_FAST               2 (result)

  3          10 SETUP_LOOP              40 (to 52)
             12 LOAD_GLOBAL              2 (range)
             14 LOAD_GLOBAL              3 (len)
             16 LOAD_FAST                0 (a)
             18 CALL_FUNCTION            1
             20 CALL_FUNCTION            1
             22 GET_ITER
        >>   24 FOR_ITER                24 (to 50)
             26 STORE_FAST               3 (i)

  4          28 LOAD_FAST                0 (a)
             30 LOAD_FAST                3 (i)
             32 BINARY_SUBSCR
             34 LOAD_FAST                1 (b)
             36 LOAD_FAST                3 (i)
             38 BINARY_SUBSCR
             40 BINARY_SUBTRACT
             42 LOAD_FAST                2 (result)
             44 LOAD_FAST                3 (i)
             46 STORE_SUBSCR
             48 JUMP_ABSOLUTE           24
        >>   50 POP_BLOCK

  5     >>   52 LOAD_FAST                2 (result)
             54 RETURN_VALUE

Code optimization

To optimize Python code, Numba takes a bytecode from a provided function and runs a set of analyzers on it. Python bytecode contains a sequence of small and simple instructions, so it's possible to reconstruct function's logic from a bytecode without using source code from Python implementation. The process of conversion involves many stages, but as a result, Numba translates Python bytecode to LLVM intermediate representation (IR).

Note that LLVM IR is a low-level programming language, which is similar to assembler syntax and has nothing to do with Python.

Numba modes

The are two modes in Numba: nopython and object. The former doesn't use Python runtime and produces native code without Python dependencies. The native code is statically typed and runs very fast. Whereas the object mode uses Python objects and Python C API, which often does not give significant speed improvements. In both cases, Python code is compiled using LLVM.

What is LLVM?

LLVM is a compiler, that takes a special intermediate representation (IR) of the code and compiles it down to native (machine) code. The process of compiling involves a lot of additional passes in which the compiler optimizes IR. LLVM toolchain is very good at optimizing IR, so not only it compiles code for Numba, but also optimizes it.

The whole system roughly looks as follows: Python numba architecture

Advantages of Numba:

  • Ease of use
  • Automatic parallelization
  • Support for numpy operations and objects
  • GPU support

Disadvantages of Numba:

  • Many layers of abstraction make it very hard to debug and optimize
  • There is no way to interact with Python and its modules in nopython mode
  • Limited support for classes


Instead of analyzing bytecode and generating IR, Cython uses a superset of Python syntax which later translates to C code. When working with Cython, you basically writing C code with high-level Python syntax.

In Cython, you usually don't have to worry about Python wrappers and low-level API calls, because all interactions are automatically expanded to a proper C code.

Unlike Numba, all Cython code should be separated from regular Python code in special files. Cython parses and translates such files to C code and then compiles it using provided C compiler (e.g. gcc).

Python code is already valid Cython code.

def sum_sequence_cython(a, b):
    result = np.zeros_like(a)
    for i in range(len(a)):
        result[i] = a[i] - b[i]
    return result

However, typed version works a lot faster.

cimport numpy as np

cpdef sum_sequence_cython(np.ndarray[np.int64_t, ndim=1] a, np.ndarray[np.int64_t, ndim=1] b):
    cdef int N = a.shape[0]
    cdef np.ndarray[np.int64_t, ndim=1] result = np.zeros([N],

    for i in range(N):
        result[i] = a[i] - b[i]
    return result
>>> timeit.timeit('sum_sequence_untyped(a, b)', globals=globals(), number=1)
Untyped Cython version 2.0215444170171395
>>> timeit.timeit('sum_sequence_cython(a, b)', globals=globals(), number=1)
Typed Cython version 0.046073039297712967

Writing fast Cython code requires an understanding of C and Python internals. If you know C, your Cython code can run as fast as C code.

Advantages of Cython:

  • Control over Python API usage
  • Easy interfacing with C/C++ libraries and C/C++ code
  • Parallel execution support
  • Support for Python classes, which gives object-oriented features in C

Disadvantages of Cython:

  • Learning curve
  • Requires expertise both in C and Python internals
  • Inconvenient organization of modules

Numba vs Cython

Personally, I prefer Numba for small projects and ETL experiments. You can always plug it into existing projects. If I need to start a big project or write a wrapper for a C library, I will go with Cython, because it gives you more control and easier to debug.

Also, Cython is the standard for many libraries such as pandas, scikit-learn, scipy, Spacy, gensim, and lxml.

If you have any questions, feel free to ask them via e-mail displayed in the footer.


  • Jameel Kassouri 2018-04-18 #

    Still unclear on one thing, if numba's object mode "often does not give significant speed improvements", why have it at all?


    • Artem 2018-04-18 #

      Object mode can be useful when you have a lot of nested loops. It gives 10-50% speedup by just adding jit decorator.