A850: Introduction to HPC for astronomers
This course is designed to give you an introduction to both n-body simulation and high-performance computing tools more generally. This course will be a lab-based hands-on exploration using sample code to illustrate the main concepts in parallel processing and use professional n-body codes
Participants will run and analyze n-body simulations in teams using the MIT Supercloud and present their results and findings during weekly meetings throughout the semester. We will explore common pitfalls, best practices, and address problems and issues together as they arise.
While this course will not require code development per se, students should be prepared to exercise all the tools of HPC, including minor changes to and compiling C (and possible C++) code, writing analysis scripts and learning about HPC job scheduling and queuing. Much of your analysis can be performed using Python and I encourage your use of Jupyter notebooks.
- The usual graduate core curriculum (esp. A645)
- Basic astronomical background (e.g. A640, A650)
- Some knowledge of numerical methods (e.g. A732)
- Self-study from recommended reading and other references
- The main work will be lab exercises that will be presented in groups every week or so. During the last month or so, we will all work on a longer project using the tools we have developed and you will present your findings during the last week of the semester.
- Grade is based on:
Performance on lab exercises and in-class presentations. 65% Performance on final project and presentation. 35%