When talking about improving Python execution performance, especially for data processing, there are too many third party libraries that may also help us. If we take into consideration their mechanisms, most of them depend on optimising the information structure or utilisation of the memory to attain performance improvement.
For instance, Dask leverages parallel computing and memory optimisation, Pandas relies on vectorisation of the dataset and Modlin optimises the utilisation of the multi-cores CPU and memory, too.
In this text, I won’t introduce any libraries. Actually, there may be a native Python decoration that may very well be used to enhance the performance significantly. We don’t must install anything since it’s built-in to Python. After all, it won’t be used for all of the scenarios. So, within the last section, I’ll also discuss when we should always not use it.
Let’s start with a vanilla example that we’re all acquainted with, the Fibonacci sequence. Below is a standard implementation using recursion.
def fibonacci(n):
if n < 2:
return n
return fibonacci(n-1) + fibonacci(n-2)
Like most other programming languages, Python also needs to construct a “stack” for a recursive function and calculate the worth on every stack.
Nonetheless, the “cache” decoration will improve the performance significantly. Also, it isn’t difficult to achieve this. We just must import it from the functools module, after which add the decoration to the function.
from functools import cache@cache
def fibonacci_cached(n):
if n < 2:
return n
return fibonacci_cached(n-1) + fibonacci_cached(n-2)
Listed below are the running results and the performance comparison.