#reed-solomon #codec #erasure #algorithm #encoding

reed-solomon-novelpoly

An implementation of a reed solomon code / encoder / decoder with complexity O(n lg(n))

6 releases (stable)

2.0.0 Jan 25, 2024
1.0.2 Sep 17, 2023
1.0.1 Aug 30, 2023
1.0.0 Mar 26, 2021
0.0.3 Mar 19, 2021

#64 in Algorithms

Download history 42975/week @ 2024-08-12 26946/week @ 2024-08-19 36711/week @ 2024-08-26 42374/week @ 2024-09-02 35508/week @ 2024-09-09 34695/week @ 2024-09-16 35341/week @ 2024-09-23 38643/week @ 2024-09-30 39885/week @ 2024-10-07 38996/week @ 2024-10-14 42216/week @ 2024-10-21 32197/week @ 2024-10-28 45913/week @ 2024-11-04 34985/week @ 2024-11-11 33585/week @ 2024-11-18 35018/week @ 2024-11-25

150,309 downloads per month
Used in 36 crates (3 directly)

Apache-2.0 AND MIT

110KB
2.5K SLoC

Rust 2.5K SLoC // 0.1% comments C 269 SLoC // 0.1% comments

reed-solomon-novelpoly

An implementation of Novel Polynomial Basis and its Application to Reed-Solomon Erasure Codes 1 2 3.

Runs encoding and reconstruction in O(n lg(n)). Note that for small number n there is a static offset due to a walsh transform over the full domain in reconstruction.

Goals

Be really fast for n > 100.

Benchmarking

For benchmarking the implementation against itself and the naive implementation, criterion is used.

cargo bench

Fuzzing

Currently honggfuzz is used.

Install cargo install cargo-hongg and run with:

cargo-hongg fuzz --bin <binary_name>

Dependencies

~0.8–1.9MB
~35K SLoC