How To Use Timeit To Profile Python Codes

May 25, 2023

Python prioritizes accessibility, readability, & convenience over performance. But that doesn’t imply you must code at a snail’s pace. Luckily, you can tweak the performance, or in other words, boost compilation, by Python’s built-in modules like timeit. Python timeit lets you evaluate the performance of functions and code on the go and decide the faster way forward.

Table of Contents

So, let’s dive in and explore the true power and functionalities of Python’s timeit and how you can use the module to make your code perform better.

What is Python timeit?

Python timeit is a simple tool to measure the processing speed or performance of small code snippets or a function. The module runs the code times a million and reports its performance. Evaluating two or three alternatives side-by-side helps make informed decisions and decide the best approach to code.

For example, looping is a common app development bottleneck, as it consumes too much time running the function through iterations. With timeit, you can compare a built-in loop with a manually written one and choose the faster one.

Python timeit Example

def f1():

      for n in range(100):

         pass

def f2():

   n=0

   while n<100:

   n+=1



if__name__=="__main__":

   import timeit

  print (timeit.timeit(f1, number=100000)

  print (timeit.timeit(f2, number=100000)

Output:

0.9181226771324873

4.5914449924603105

The above program compares two ways to run 100 iterations; Python’s built-in function (f1) and increments (f2). As specified, timeit modules run both 100,000 times and share a total runtime.

Remember, the module uses a million runs by default though you can specify and bring down the count as shown in the example above. Undoubtedly, a million runs get you better results though a little less gets you workable results faster.

Using Python’s timeit: The String Approach

The method described above is not the only way to run the timeit module, as you can also do so by passing a string.

import timeit

print (timeit.timeit('for n in range(100):pass'))

The above is the usual way, while below is the command line to run the module via string.

python -m timeit "for n in range(100): pass"

In comparison, the first approach is relatively better as it saves you from forcing the command via text string.

When To Not Use Python timeit?

Undoubtedly, Python timeit is a timesaver and can help you write faster codes in Python and up your efficiency. However, using the module to profile the entire program is not recommended. Even though it is not against the rule to profile a program with Python’s timeit, it is not the best way.

The module is best for profiling snippets or a few lines, and other modules like cProfile are relatively better for complete programs. Firstly, the latter gets you more detailed statistics and helps make program-relevant decisions. Secondly, it is more reliable.

That said, profiling the entire program with the timeit module will still get the results, but it would be inconsistent and not as reliable as cProfile’s verdict.

Moreover, the Python timeit module is designed for short and quick functions to help developers make them even quicker and not so for functions that take a while to process, as it beats the purpose.

Python timeit: Tips To Remember

Before doubting Python’s timeit module efficiency, it is essential to understand programs hardly ever run the same way. The performance and speed rely on many factors & uncertainties like resource allocation, cache behavior, priorities, etc. For instance, running a program with multiple applications opened in the background won’t be as quick as running it individually.

For the same reason, timeit runs a code through a million iterations. Still, the act doesn’t ensure 100% efficiency, and it is best to run aggregate trials to get as near to real performance speed as possible. The best trick is to run timeit modules multiple times, ditch excessively high and low scores, and average the meaningful ones.

Moreover, it is best to run the module on different machines or systems for even better results. The more tests you run, the higher accuracy you achieve. However, it is essential to know you must balance the efforts. Just like it is rash not to test approaches with modules like Python timeit, it is imprudent to get stuck and run excessive tests in the quest for unachievable perfection.

Also Read: How To Add Items To a Python Dictionary?