Course Overview

## Welcome Penn State Astro 528 High-Performance Scientific Computing for Astrophysics Eric Ford --- ## Safety & Respect ___ ### Masks - As of August 4, all students, faculty, staff and visitors are required to wear masks at all times in all indoor public spaces, including classrooms. - Any students who choose to join class from Davey 538 are expected to wear masks properly, covering the bridge of their nose and under their chin. - Studies indicate that surgical masks are much more effective than cotton masks. ___ ### Physical Distancing - While some students may be comfortable working closely with a partner, others may prefer to maintain physical distance. - Students are expected to respect others' requests for physical distancing. - It's ok to change one's mind mid-semester, as we learn more about the situation. ___ ### Please err on the side of caution - If not feeling well, please stay home. - If have reason to believe may have been exposed, please stay home. ___ - Being extra cautious during class will help peers and instructor to focus on learning, rather than worrying about their safety. - If engage in less safe activities outside of class, please take precautions. --- ## Course goals Enhance your skills for scientific computing - Increase your productivity + Choose right tool for right task + Reduce time debugging + Improve reproducibility - Enable you to + Analyze "Big Data" + Increase resolution of simulations + Include more complex physics + Increase impact of your software --- ### Course outline - Software Development Practices - Writing efficient serial code - Parallelizing code efficiently ___ ### Software Development Practices Note: Ask what students think of when they hear "software development practices". ___ ### Software Development Practices - Version control - Testing & Continuous Integration - Debugging - Documenting & Literate Programming - Coding standards - Reviewing code - Reproducibility - Workflow ___ ### Writing efficient serial code - Processor architectures - Memory hierarchy - Networking - Programming languages - Choosing algorithms - Benchmarking - Profiling - Compiler optimizations - Optimizing ___ ### Parallelizing code efficiently - Shared-memory (e.g., one workstation) - Distributed-memory (e.g., cluster) - Accelerators + GPUs + TPUs (volunteers?) - Cloud Note: Ask if any students already using parallel codes. If so, how were they parallelized? --- ### Specific Objectives - Increase technical knowledge + Readings, online lessons & class discussion - Practice fundamentals on a small scale + Lab/homework exercises + Make lots of mistakes quickly & learn from them + Make good habits routine - Transfer skills into real work environment + Class project + Apply new skills to your research + Build deeper expertise in topics most relevant to you + Share what you learn with the class ___ ### Readings, presentation ![textbooks](/images/textbooks.jpg) ___ ### Readings, presentation - Textbooks + _Writing Scientific Software: A Guide to Good Style_ + _ThinkJulia: How to Think like a Computer Scientist_ + _Introduction to High Performance Computing for Scientists and Engineers_ (definitely optional) - Online PDFs - Online tutorials - Recordings (will try, no guaranty) Note: Ask students if they would like to record classes. Ask students if they like the idea of a pre-recorded lesson prior to class. ___ ### Class discussions - Let's learn from each other - Introductions + Name (& pronouns if you like) + Department (if not Astro) + Year of your program + What skill you hope to strengthen through this course ---
## Let's get you set up ### Accounts - Penn State - [Roar (ICDS-ACI/Roar)](https://ics.psu.edu/computing-services/account-setup/) - [Github](https://github.com) - [TopHat](https://www.tophat.psu.edu/)
___
### Get Started For your first lab session, you'll: - Follow link for lab 1 from Canvas announcement - View _your_ new repository on [Github](https://github.com) - Login to where you'll run your code + [ICDS-ACI Portal](http://portal.aci.ics.psu.edu/) + [Install locally](https://julialang.org/downloads) (not for today) - _Clone_ your "repo" for the lab - Start Pluto notebook server - Start working through ex1.jl, then ex2.jl, ... There are [more detailed instructions](/tips) on the website. ___ ### Commiting Changes - [_Commit_ your changes to local "repo"](https://psuastro528.github.io/tips/submitting/commit/) - Can commit Pluto notebooks directly as .jl files. - Later one when using Jupyter notebooks + First make a markdown version for human-readable version control + Use Weave's convert_doc to convert Jupyter notebook (ex1.ipynb) into Julia Markdown (ex1.jmd) + Add & commit both ex?.ipynb and ex?.jmd files to your local repo - Before signing off for the session, ["push" your commits to GitHub](https://psuastro528.github.io/tips/submitting/submitting/#commit-push) - See [more detailed instructions](/tips/submitting) & ask questions as you go ___
# Questions? [Jump to start](#/0/0")
  • See it fullscreen