How To Make Your Code Faster
Computing and programming is now becoming an indispensable part of statistical research, however we often run into a situation that the computation time is intolerably long for simulation studies. There are many factors that may influence the efficiency of the code, for instance the algorithm for computing, the tools/packages you use, and also the programming languages.
In this talk I’ll share some of my own experience in accelerating the code for statistical computing. Three major aspects will be discussed:
- How to choose proper numerical algorithms and techniques to avoid unnecessary computation cost.
- Introducing and recommending useful packages for high performance computing, for example how to read and summarize data efficiently, how to run fast linear models, etc.
- I’ll also introduce some languages other than R that may help writing faster code, for example Julia and C++/Rcpp.
Seminar Information
- Speaker: Yixuan Qiu
- Date: Thursday, February 11, 2016
- Time: 4:30 pm - 5:20 pm
- Location: Math Library Lounge