Rust Outperforms Python in Quicksort | Generated by AI
Results: Rust ~0.75ms vs Python ~7.9ms — ~10x faster.
Here’s the post:
Quicksort in Rust vs Python: sorting 10,000 numbers
I implemented the same Quicksort algorithm from scratch in both Rust and Python — identical Lomuto partition logic — and timed them on 10,000 random integers.
Results:
- Rust (release build): ~0.75 ms
- Python: ~7.9 ms
That’s roughly 10x faster for the same algorithm, same logic, no tricks.
The code is nearly a 1:1 translation. Python’s tuple swap (a, b = b, a) is arguably more elegant than Rust’s arr.swap(i, j), but Rust compiles down to raw machine code while Python is interpreted with dynamic typing overhead at every step.
What’s interesting is the gap is “only” 10x here. Python gets a bad reputation but for algorithmic work on moderate data sizes, it’s not catastrophic. Push to 1M numbers or do this in a hot loop and the gap widens fast.
The real lesson: the algorithm matters more than the language at small scale. Both are O(n log n) average case. But when you need every millisecond, Rust’s zero-cost abstractions deliver.
Code: src/main.rs (Rust) and quicksort.py (Python) — same algorithm, very different runtimes.
Want me to commit this and push?