Summary

Implemented functions that operate on matrices and vectors in RISC-V assembly, and used them to construct a simple Artificial Neural Net (ANN) that is able to classify handwritten digits to their actual number. Not super complex, but demonstrates extremely useful systems can be created with only basic, low-level tools.

I learned a lot from doing this, including:

• how to use registers efficiently
• calling convention
• allocating memory on the stack and heap
• working with pointers