What's the main reason ripgrep is fast?

According to the repository data, ripgrep's speed comes from multiple factors working together:

The primary reason is that it's built on top of Rust's regex engine, which uses finite automata, SIMD, and aggressive literal optimizations [README.md:L187-L189]. Rust's regex library also maintains performance while supporting full Unicode by building UTF-8 decoding directly into its deterministic finite automaton engine [README.md:L190-L192].

Beyond the regex engine itself, ripgrep employs several other performance strategies:

1. Smart search strategy selection: It supports searching with either memory maps (better for single files) or incremental buffering (better for large directories), and automatically chooses the best approach [README.md:L193-L196].

2. Efficient ignore pattern matching: It applies .gitignore patterns using a RegexSet, allowing a single file path to be matched against multiple glob patterns simultaneously [README.md:L197-L200].

3. Parallel directory traversal: It uses a lock-free parallel recursive directory iterator via crossbeam and ignore libraries [README.md:L201-L203].

The README also notes that ripgrep supports optional PCRE2 regex engine switching via the -P/--pcre2 flag [README.md:L149-L152], though the default regex engine is tuned for speed.

FOLLOW_UPS:

• How does ripgrep handle UTF-8 decoding performance?
• What's the difference between memory maps and buffering?
• Can you disable automatic filtering for speed?