---

name: performance description: Analyze and optimize code performance with profiling and benchmarking. Use when asked to improve performance or diagnose slowness.

Performance Expert

Analyze and optimize code performance.

Performance Analysis Process

1. Measure - Profile before optimizing
2. Identify - Find the actual bottleneck
3. Optimize - Target the hot path
4. Verify - Measure improvement

Common Performance Issues

Memory

• Unnecessary cloning
• Large stack allocations
• Unbounded collections
• Memory leaks (Rc cycles)

CPU

• Inefficient algorithms (O(n²) when O(n) possible)
• Repeated computations (missing memoization)
• Excessive allocations in hot paths
• Blocking async runtime

I/O

• Synchronous I/O in async code
• Missing buffering
• Too many small writes
• N+1 query patterns

Rust-Specific Optimizations

Avoid Unnecessary Clones

Copy// ❌ Bad
let x = expensive.clone();

// ✅ Good: Borrow if possible
let x = &expensive;

// ✅ Good: Move if not needed after
let x = expensive;

Use Iterators

Copy// ❌ Bad: Intermediate allocations
let v: Vec<_> = items.iter().map(f).collect();
let result: Vec<_> = v.iter().filter(g).collect();

// ✅ Good: Chained iterators
let result: Vec<_> = items.iter().map(f).filter(g).collect();

Async Best Practices

Copy// ❌ Bad: Blocking in async
async fn bad() {
    std::thread::sleep(Duration::from_secs(1));
}

// ✅ Good: Async sleep
async fn good() {
    tokio::time::sleep(Duration::from_secs(1)).await;
}

Profiling Tools

Copy# CPU profiling
cargo flamegraph

# Memory profiling
cargo +nightly miri

# Benchmarking
cargo bench

# Check binary size
cargo bloat --release