Measured - byte-exact - 2026

Pedulli vs LZFSE

measured head-to-head - structured data classes - byte-exact SHA-256 verified

TL;DR - honest
Apple LZFSE targets Apple-platform throughput at a zstd-level-3-ish ratio. Pedulli is a best-of-N racer - it races xz, zstd, brotli and your data's SRD math, keeps the smallest verified output, so it is never larger than the best standard codec (worst case +1 byte). It wins outright on structured data and ties the best codec on already-optimal/random data; LZFSE wins on macOS/iOS native integration and energy efficiency on Apple silicon. Different design targets.

The measured table

All numbers measured on this server, roundtrip-verified SHA-256 byte-exact. Proofs available on request.

InputLZFSEPedulli (best-of-N)Δ
1 MiB of zeros692 B13 Bmuch smaller (redundant input)
JSON 31 KB3,894 B1,265 B-68%
HTML 161 KB49,210 B38,757 B-21.2%
Apache logs 3.5 MB289 KB217 KB-24.9%
MP4 master 10 MB9,651,123 B9,430,108 B-221 KB
Random bytes 1 MB1,048,608 B (+32 B)1,048,577 B (+1 B)31 B less overhead

What LZFSE does better (honest)

What Pedulli does that LZFSE does not

When to use which

Use LZFSE for iOS/macOS native file IO where power matters more than ratio. Use Pedulli when storage cost compounds across a large corpus regardless of platform - plus the per-file +1 byte never-worse floor.

More comparisons: lz4 - zstd - Snappy - LZO