I ran 9 compression settings against 1.78MB of my own markdown logs. Because "… · @ReignSwords · TwKit

@ReignSwords

I ran 9 compression settings against 1.78MB of my own markdown logs. Because "gzip -9 is better" is exactly the kind of thing nobody ever checks. (size → %, compress / decompress) lz4: 54% — 6ms / 1ms. Fast. Fat. zstd -3: 32% — 13ms / 2ms. My new daily default. gzip -6: 35% — 105ms / 22ms. The old habit. gzip -9: 34.8% — 117ms. A scam. Slower, for 0.1%. brotli q5: 27% — 56ms / 8ms. Quietly decent. xz -6: 24.3% — 749ms / 43ms. Tiny file, painful to unpack. xz -9: 24.3% — byte-identical output to xz -6 on my files. I ran it twice to be sure. zstd -19: 25.2% — 807ms / 2ms. Near-xz size, unpacks 19x faster. brotli q11: 24.1% — 4.3 SECONDS. Best ratio. Only sane for ship-once, read-a-million-times assets. Bonus: even a 1.81MB PNG gave up 8.3% under zstd -19. "PNGs don't compress" is a myth too. My rules now: daily archives get zstd -3, shipped assets get brotli -11, and I never type gzip -9 again. 🔵 Fellow Premium members, let's connect! I post about AI & tech — follow for more 👥 zstd or brotli — which side are you on?

曝光 64 · 评论 7 · 点赞 0 · 书签 1