news
Programming Leftovers
-
Martin Chang ☛ Adventuring into Tenstorrent SFPU Programming (Mandelbrot rendering)
First a recap on what the Mandelbrot set is. For our purposes, the Mandelbrot set is a set of complex numbers c for which the function f(z) = z^2 + c does not diverge when iterated from z = 0. That is, for all n in the natural numbers, the absolute value of f^n(0) remains bounded. And we can color the points in the complex plane based on how quickly the sequence diverges. For example, if the sequence does not diverge, we can color the point black. If the sequence diverges, we can color the point based on the iteration count.
-
Wired ☛ The Subjective Charms of Objective-C
Objective-C came up in the frenzied early days of the object-oriented programming era, and by all accounts, it should have never survived past it. By the 1980s, software projects had grown too large for one person, or even one team, to develop alone. To make collaboration easier, Xerox PARC computer scientist Alan Kay had created object-oriented programming—a paradigm that organized code into reusable “objects” that interact by sending each other “messages.” For instance, a programmer could build a Timer object that could receive messages like start, stop, and readTime. These objects could then be reused across different software programs. In the 1980s, excitement about object-oriented programming was so high that a new language was coming out every few months, and computer scientists argued that we were on the precipice of a “software industrial revolution.”
-
GCC ☛ GCC 15.1 Released
Driving a leading free software project such as GCC would not be possible without support from its many contributors. Not only its developers, but especially its regular testers and users which contribute to its high quality. The list of individuals is too large to thank individually!
-
Rlang ☛ ‘Bayesian’ optimization of hyperparameters in a R machine learning model using the bayesianrvfl package
The surrogate model used for Bayesian optimization is a Non-Bayesian Gaussian Random Vector Functional Link (RVFL) network (instead of a Gaussian Process) (see Chapter 6), whose number of nodes in the hidden layer and volatility of residuals are chosen by using maximum likelihood estimation (MLE). This surrogate model is trained on 10 results of the objective function evaluations, and an Expected Improvement acquisition function is used to determine the next point to sample in the hyperparameter space.
-
Perl / Raku