Artagnon's Presence

A couple of my talks have been shortlisted, but the abstracts won’t be up on the site until a week before the conference. So here are the abstracts. I’ve just started preparing the talks, so feedback is appreciated.

An Insight into CPython Compiler Design

The objective of this talk is to give the audience an insight into compiler design, specifically the design of the CPython compiler. The talk attempts to do this by taking apart the Unladen Swallow codebase.

Starting from a small hand-waving discussion using block diagrams, we will discuss compiler optimization techniques introduced by Unladen Swallow using revision control to playback changes. There is (and will be) no original research; only implementation of ideas from existing research papers.

A significant portion of the talk will be dedicated to compiling Python bytecode to LLVM IR (in Python/llvm_compile.cc), and the eval loop (in Python/eval.cc). While there are thousands of possible optimizations, only those which yield significant benefits and are in line with CPython’s objectives will be discussed.

Intended audience: Programmers interested in understanding (and ultimately contributing to) compiler projects like Clojure, Unladen Swallow, GHC.

Prerequisites: Requires a basic understanding of language grammers and parsing [1], as well as elementary compiler design [2]. The talk will jump right into the codebase with little introductory material, so the audience is required to do a little background reading.

[1] Flip through the relevant parts of the Dragon Book
[2] Scheme’s eval-apply loop from SICP should do fine

Haskell Internals

The session will focus on writing clever code in non-monadic Haskell, by solving several problems modeled on problems in [http://projecteuler.net] and [code.google.com/codejam/]. It will serve as an introduction to folds, maps, laziness, lexical closures, and other core features of functional programming in Haskell, while also attempt to give a small peek into the working of the Glasgow Haskell Compiler.

Target audience: People interested in functional programming and people interested in scientific computing.

Prerequisites: Ideally none. But since  the session intends to have little introductory material, a quick read through some jargon will be necessary.

curve-fitting.pdf (168 KB)

I delivered a talk on curve fitting today in the physics department seminar room. This is the presentation I used. The objective of the talk was to familiarize everyone with the chi-square distribution and chi-square fitting without focusing too much on the implementation details. Since the complete audience didn’t have background knowledge of probability theory, I had to briefly touch up some concepts like Cumulative Distribution Function and Probability Density Function.

Overall, I think the talk came out alright, although it was apparent that some of my fundamentals were holed. It’s a miracle it even came out, considering that I was delivering it on a paracetamol-suppressed headache and fever.

Finally, LaTeX + Beamer is too awesome!

Edits: The zeroth central moment is one, not zero.