This program started out as a Python script, which generally worked fine.
It would however crash every now and then after running for a long time. In an effort to find what was causing that I started logging exceptions to syslog. This did not result in any clues.
It is a bit silly to have a complete python interpreter running to produce a statusline. So I decided to rewrite it in C.
To learn how it works and because I don’t use Java, I decided to port professor Michael Pound’s implementation of the Enigma machine and the fitness and analysis code.
The initial port was done to Python. Unfortunately, this is one instance where Python was rather slow because of the large amount of settings that need to be tried.
So I decided to port it to C instead.
This is the fifth in a series of articles that covers analyzing and improving performance bottlenecks in Python scripts. The program calculix-orient is a helper program that automates the asignment of local coordinate systems for anisotropic finite elements. It is designed to work together with CalculiX.
After having used the built-in tracing profiler cProfile for a while now, I came across the statistical (sampling) profiler pyinstrument.
A statistical profiler does not influence the runtime of the program as much
as a tracing profiler as cProfile does.
Additionally, pyinstrument shows the result in a different way from cProfile.
So in this article, I will be profiling my repotool program.
This program helps me maintain a mirror of FreeBSD packages that I use.
This article explains how to make CalculiX nodes and elements accessible with Python.
Recently I saw that you could convert python scripts into executable with the
help of cython, and I had to try that.
As an engineer, I write a lot of small python programs as tools for specific tasks. Generally, these are not large programs. Most of them are below 100 lines of code (“LOC”, as measured by cloc), although there are a few in the 300−400 LOC range.
In this article, I will present some observations about these, and draw some conclusion from them.
Since Python 3.11.0 came out recently, I wanted to compare its speed with 3.9.15 on some of my own “benchmark” programs.
Both versions are in the default configuration as built by the FreeBSD ports system.
TL;DR Python 3.11 yielded speed improvements in the order of 25% in my tests.
In this article we will discuss a faster and more efficient way to get only the first item from a view object in Python.
Intrigued by this question, I investigated the structure of an OpenSSH ECDSA public key.