- **rolled for loops**: the cost of parsing unrolled loops grows linearly, so currently your compile times actually grow with higher iteration counts :( - **mutable references**
