This is a small benchmark to compare Serde and Nanoserde at compilation speed, and the difference is a lot, like a lot a lot! In my PC (MacBook Pro M1 16GB RAM 512GB SSD) here are the results:
./build-test.shMain code is written in ./benchmark. Best thing you can change easily is the function_count. This sets the number of functions that derive serde/nanoserde serialize and deserialize. Build times depend on this 'lil usize value.