AMAZINGINDEX.COM 日报快照
60.6
VOL. 2026.07
2026.07.16
← 返回 2026.07.16 日报
日报快照 · Daily Snapshot
NO. 012

新编解码器解压速度超LZ4两倍

#ARTICLE HackerNews 2026.07.16
推荐指数 72.0 NO. 012 · 2026.07.16
发布2026/07/15Score78Comments28

misa77是一款面向写一次读多次场景的LZ系编解码器,解压速度比LZ4快2倍且压缩比更优,压缩时内存占用不超过5MB、解压时零内存分配。对于需要高频读取冷数据或带宽受限的AI推理服务,这能直接降低存储IO瓶颈和内存压力。

AI模型权重和中间结果的存储成本被严重低估。当前主流做法是直接用LZ4或Snappy做快速压缩,但推理服务加载大模型时解压环节常成为冷启动瓶颈。misa77的零内存解压特性对GPU显存紧张的场景尤其有价值,比如边缘设备部署或Serverless推理。

压缩速度慢的问题在模型发布场景下几乎不构成约束——权重文件生成一次后要被下载数百万次。HuggingFace、模型仓库和MLOps平台是最该评估这套方案的客户,个人开发者也可以关注其是否会被集成到safetensors或gguf等模型格式中。

意见分歧 27 条评论

核心争论:解压速度提升是否值得牺牲安全性与跨平台一致性

bootlegbilly

this is super interesting! im excited to give this a look this afternoon, since I specifically have wanted faster throughout for decompressing maps in a game engine.

nonadhocproblem

Hey, glad to hear that you found it interesting. misa77 primarily targets textual data (ie. byte-aligned data formats where each byte corresponds to a symbol), so I hadn't tested it on game assets much until now. After seeing your comment, I pulled some random assets from Pathfinder WoTR (in fact, U

zX41ZdbW

You can look at the LZ4 decompression implementation in ClickHouse: https://presentations.clickhouse.com/2018-highload-siberia/ (the presentation is quite old, but the implementation was updated recently). It uses the same LZ4 format, just decompresses faster.

替代方案: LZ4OodleSelkiezxc
查看原文 →