SHOGUN is a machine learning toolbox whose focus is on large scale kernel methods and especially on Support Vector Machines (SVM). It provides a generic SVM object interfacing to several different SVM implementations, all making use of the same underlying, efficient kernel implementations. Apart from SVMs and regression, SHOGUN also features a number of linear methods like Linear Discriminant Analysis (LDA), Linear Programming Machine (LPM), (Kernel) Perceptrons, and algorithms to train hidden Markov models. SHOGUN can be used from within C++, Matlab, R, Octave, and Python.
Loadbars is a small script that can be used to observe CPU loads of several remote servers at once in real time. It connects with SSH (using SSH public/private key auth) to several servers at once and vizualizes all server CPUs and memory statistics right next each other (either summarized or each core separately). Loadbars is not a tool for collecting CPU loads and drawing graphs for later analysis. However, since such tools require a significant amount of time before producing results, Loadbars lets you observe the current state immediately. Loadbars does not remember or record any load information. It just shows the current CPU usages like top or vmstat does.
xBaK is a fan-made remake of the classic Sierra computer RPG "Betrayal at Krondor". It is a game engine that uses the data files that came with the original game by Sierra Online. You must already have your own copy, since the required data files are not distributed with xBaK. The game is still under development. The intro, the option dialogs, and the main game dialog are finished, but the game is not yet playable. Several tools for examining the contents of the data files are also available.
ArduinoPulseGenerator is a simple program for generating pulse sequences using an Arduino. There is an associated GUI that runs on the local computer, or you can simply connect to the Arduino with a serial console (9600 baud) and send it commands. This code has been tested on the ArduinoMega 2560 (timing accuracy ~ ±200 μs) and Arduino Due (timing accuracy ~ ±35 μs); it may work on other Arduino boards.