Why Write Your Own Hydro Code?

There are a lot of simulation codes developed by the astro community that are freely available (and you are hearing about some during this school).

So why should you write your own?

Tip

My advice is to write your own implementation of the type of solver you’ll be using for your research once, even if you don’t use it for your actual work.

The process of figuring out how to organize and write the code will demystify a lot of what takes place in a modern simulation code.

It’s difficult to completely document a code in a paper

There are lots of subtleties

Algorithms for (magneto-,radiation,relativistic-)hydro are complex (often 10s of thousands of lines of code). While we do our best to document the algorithm in a paper, there are so many small implementation details that won’t be in the paper (either for space or simply because they are overlooked by the authors).

This means that the only real documentation for a simulation code is the code itself.

If you want to truly understand how an algorithm works, coding it up yourself and implementing tests is incredibly useful.

There are strengths / trade-offs that need to be made along the way

All algorithms make approximations: from the choice of equations solved to the method of discretizing them in space and time, and to the details of the additional physics included.

A great way to understand where these trade-offs enter is to force yourself to have to make these decisions in your own code.

Eventually you’ll want new features

If you are using an established research code, eventually you’ll want to add new features and contribute to the codebase. Again, these codebases are large, and the complexity of modern supercomputer architectures can sometimes make the core methods more opaque.

Having a clear understanding of how to implement your method in a simple fashion on your own will help you understand the large research codes that our community uses.

Our community continually needs new developers

Research codes are continually improved and this is done by members of the community, usually out in the open, e.g., via github. Going through the experience of putting together your own code prepares you for the development that is needed for community codes. It also gives you a simple well-understood testbed to explore new ideas before proposing them in a community code.