Events

The Sable group does several presentations and workshops throughout the year. These events are open to all interested McGill students, staff and faculty. Anyone outside of McGill should contact Dr. Hendren in order to arrange attendance to any interested event.


  • McNumJS

    Date: September 16, 2014
    Time:12 Noon
    Location: MC320 McConnell
    Abstract:

    McNumJS is a numerical library for JavaScript which provides NumPy like API and it uses recent JavaScript features like Typed Arrays to make it fast. The presentation will discuss the development and architecture of the library as well as some initial performance results. Anyone interested in developing numerical application in JavaScript, will find this talk resourceful.


  • Git

    Date: September 23, 2014
    Time:12 noon
    Location: MC320 McConnell
    Abstract:

    Git is a version control system that has been gaining in popularity, but it has a reputation for a steep learning curve and an intimidating command line interface. In order to avoid having to memorize arcane incantations to make git do what we want, it can be beneficial to learn to reason about how git actually works. This presentation will be in equal parts theoretical and practical. To keep things computer sciencey we'll talk about git's underlying data model and how git does what it does, but there will be illustrative live demos throughout, so if you've never used git, you should walk away with a working knowledge of how the most common commands are used, in addition to a high-level picture of how the whole thing works.


  • VeloCty

    Date: September 30, 2014
    Time:12 noon
    Location: MC320 McConnell
    Abstract:

    VeloCty is an optimizing static compiler for Matlab and Python and performs partial compilation of code written in these two languages to parallel C++. Matlab and Python with its NumPy library are also gaining popularity among scientists and mathematicians. These languages provide many features which allow easy prototyping. However these same features inhibit performance of the code. We developed VeloCty as a solution to the problem of enhancing performance of programs written in these languages.VeloCty takes as input, functions written in Matlab and Python which are defined by the user as computationally intensive and generates an equivalent C++ version.  The talk will focus on the  challenges of supporting the semantics of two languages, implementation and design on VeloCty and the performance improvement seen as a result of the compilation.