An introduction to approximate computing

Approximate computing is a wide spectrum of techniques that relax accuracy of computations in order to improve performance, energy, and/or other metric of merit. In this article, I’ll try to provide a structured introduction to this research area.

Undefined behavior in the wild

So I was hacking an open source C++ project. The project builds using gcc v5.4 and runs without problems. Then, I grabbed gcc v6.2 to see how much performance improvement can be gained. Surprisingly, the program segfaulted …

Floating-point satisfiability as global optimization

Solving SMT formulas involving the theory of floating-pointing arithmetic (FPA) can be encountered in several domains including symbolic execution and program synthesis. Unfortunately, SMT solvers which support FPA often struggle trying to satisfy …

Building this blog

Details about the tools and services used in building this blog. Also, some discussion about the design rationale behind it.