I always wrote a lot of code, either to evaluate whether a language or an API fits my needs, or just for fun. I tried a lot of languages. My opinion is that C++ is mandatory for large scale, performance-critical software. I mean, C also fits these requirements, however ad-hoc polymorphism and a macro jungle cannot be in par with C++ built-in object-oriented features and template meta-programming.
I would like using Ruby for scripting purposes, but the huge amount of available libraries makes Python more attractive. I like coding something in Haskell, once in a while, and I usually perform black-magic with autotools.
Here you can find a list of projects I uploaded to GitHub.
My GitHub pages site repository. As you can see in the footer, I just put some contents on top of Jekyll together with a theme from Mark Reid – thank you.
In 2012 I was teaching assistant of the Code Optimization and Transformation course, held at Politecnico di Milano by Professor Stefano Crespi Reghizzi. During the classes, I show students how to exploit LLVM libraries in order to build simple compiler analyses and transformation tools.
To ease student lifes, I prepared a sample project. It contains examples about how to define/link/test a new LLVM pass. It is a little bit outdated, but consider that it should be used by students, so working on trunk is not an option.
A long time ago, after coding a while inside GCC, I decided to try LLVM. I was impressed by the readability of the code: anyone, with basic skills about compilers and C++ can easily understand how it works, and use it some project.
Thus I decided to write something in order to learn its API. I wrote a small OpenCL runtime for multi-core machines. I can run simple OpenCL applications, but, unfortunately, I had not much time to work on it, so its development has been stopped. I really hope to find some spare time in the future to continue improving on it.
I did the Ph.D. at Dipartimento di Elettronica e Informazione – Politecnico di Milano. Unfortunately, the template for slides is available only in the PowerPoint format, so I decided to try porting it to Beamer.
Code repository is on GitHub:
Here are release tarballs and previews: