Nice!

[udif]

It's good to see another SystemRDL tool to enhance the echosystem. (though I haven't tried your tool yet).

I'm curious why have you decided to write your own tool from scratch instead of using the Python-based PeakRDL.
After all, Python development is quicker, and a SystemRDL tool doesn't nead real-time performance.

[vowstar]

Thank you for looking at the project and for the encouraging words about growing the SystemRDL ecosystem!

I decided to re-implement the library in C++ mainly to make it a drop-in runtime dependency for qsoc.
That project is written entirely in C++, so keeping everything in the same language spares us from shipping a full Python stack, fighting Python 2 vs Python 3 issues, or asking users to install Python on machines that might not have it at all.

There were also two technical gaps I kept hitting with PeakRDL:

• Arbitrary register widths – SystemRDL 2.0 allows any width, but PeakRDL requires each register to be a power of two. Some of our RTL blocks need odd sizes, and stretching them to 8/16/32 bits adds a lot of padding logic.

• Automatic reservation of unused bits – During elaboration I want all holes to be marked as reserved so back-end code generators don’t need to repeat that bookkeeping.

Building a small, standalone C++ library that solves those points and compiles alongside qsoc turned out to be the simplest path.

I hope that clarifies the motivation. I’m always happy to discuss design choices or accept patches, so feel free to open an issue if you think something can be improved.