Description:

Once you have an application that works well and supplies accurate results you may ask yourself why the run-time is slow or why it doesn't scale as you expect it to.  If your code is running in parallel using message passing or multithreading, identifying bottlenecks becomes even more complex. Fortunately, there are tools that take the guess-work out of identifying the sections of your code responsible for the bulk of your run-time.  We’ll look at some of the most used profiling tools available for R, Python, and compiled software. Profiling output will be produced, and we'll demonstrate how to interpret the results, which should guide how to adjust your code for more efficient computation. This workshop will not be recorded. 
 

Prerequisites: Before attending this workshop, researchers new to Quest should:   

• Apply for a Quest allocation. We recommend “Research Allocation I” if you are new to Quest. If your lab has a joint Quest allocation, you can also apply to join that allocation. 
• Watch the Introduction To Quest video series to get an overview of the system, learn how to submit jobs, and become familiar with best practices. To learn more about Quest, see  About Quest. 
• Become familiar with Unix command line: see our Intro Command Line and Bash Scripting workshops and check Online Resource Guide for learning basic command line skills.