Show HN: ZXC – Asymmetric, +40% decode vs. LZ4 on ARM (C, BSD-3, Fuzzed)
3 days ago
- #embedded-systems
- #performance
- #compression
- ZXC is an asymmetric high-performance lossless compression library optimized for Content Delivery and Embedded Systems.
- It focuses on 'Write Once, Read Many' scenarios, trading compression speed for maximum decompression throughput.
- ZXC outperforms LZ4 decompression by +40% on Apple Silicon and +22% on Cloud ARM (Google Axion).
- Officially merged into the lzbench master branch, allowing independent verification via industry-standard benchmarks.
- Designed for asymmetric efficiency, ZXC optimizes decompression for modern CPUs, particularly ARMv8.
- Benchmarks show ZXC -3 (Standard) is 1.39x faster than LZ4, while ZXC -5 (Compact) is 3.3x faster than Zstd-1.
- Available for multiple platforms including macOS ARM64, Linux AArch64, Linux x86_64, and Windows x86_64.
- Provides a CLI for basic compression, high compression, decompression, and benchmarking.
- Includes a thread-safe API suitable for highly concurrent environments and multiple programming languages.
- Features continuous fuzzing, static and dynamic analysis, and a safe API requiring explicit buffer capacity.