rudu is a high-performance, Rust-powered replacement for the traditional Unix du command. It uses multithreading (rayon) for fast parallel directory traversal, reports true disk usage (st_blocks × 512), and adds filtering, CSV export, incremental caching, and memory limiting — all with Rust's memory-safety guarantees.

rudu                                          # scan current directory
rudu /data --depth 2 --sort size              # top 2 levels, largest first
rudu /project --exclude .git --output out.csv # exclude .git, export to CSV
rudu /lustre --memory-limit 900 --no-cache    # HPC cluster with 1 GB job limit

For the full options reference and annotated examples, see docs/basic-usage.md.

• Parallel scanning — work-stealing thread pool via rayon; configurable with --threads N
• True disk usage — st_blocks × 512, same as du
• Depth & exclusion filtering — --depth N, --exclude PATTERN
• Flexible output — terminal table or --output report.csv
• Owner & inode info — --show-owner, --show-inodes
• Incremental caching — skips unchanged subtrees on repeat scans; --no-cache, --cache-ttl
• Memory limiting — --memory-limit MB for HPC/SLURM jobs; graceful degradation at 95 % of limit
• Performance profiling — --profile prints per-phase timing

* For very small directories, startup and threading overhead outweighs the benefit. Performance gains scale with directory size.

# From source
git clone https://github.com/greensh16/rudu.git
cd rudu
cargo build --release
cargo install --path .

# From crates.io
cargo install rudu

Contributions are welcome. For major changes please open an issue first.

cargo test
cargo fmt --check
cargo clippy --all-targets -- -D warnings
cargo doc --open

GNU General Public License — see LICENSE.

Built with Rust · parallel processing via Rayon · CLI via Clap · progress via Indicatif · memory monitoring via sysinfo