Parking_lot for faster synchronization primitives

Marcus Chen Jan 2026

1 tab

use parking_lot::Mutex;
use std::sync::Arc;
use std::thread;

fn main() {
    let counter = Arc::new(Mutex::new(0));
    let mut handles = vec![];

    for _ in 0..10 {
        let counter_clone = Arc::clone(&counter);
        handles.push(thread::spawn(move || {
            *counter_clone.lock() += 1;
        }));
    }

    for h in handles {
        h.join().unwrap();
    }
    println!("Count: {}", *counter.lock());
}

1 file · rust Explain with highlit

parking_lot provides drop-in replacements for Mutex, RwLock, and Condvar that are faster and smaller than std::sync. They use more efficient parking/unparking and don't poison on panic. I use parking_lot::Mutex in hot paths where lock contention matters. The API is nearly identical to the standard library, so switching is easy. The main difference is that parking_lot doesn't return LockResult, so you don't need .unwrap(). For production services with high concurrency, parking_lot can noticeably improve throughput. It's a simple dependency change that often pays off.