fig (aka:[phil]) @bad-example.com · Apr 11 1 / 3
gonna break these down a bit in a thread-that-should-probably-be-a-post 1. lz4 compression is bad! like, 50% worse than zstd-1 (without dictionary training) block sizes up to 128k but it's fast right? 4k blocks: ya, 35% faster 8k: no! -37% (slower!) 16k: -14% 32k: -3% 64: -1% slower 128: +4% fast
1 likes 1 replies
?
Replies
changing zstd compression level (without dictionary training) makes a surprisingly small difference. very interesting how 1/3 and 6/9 start bunched, then fan out evenly as data increases. i guess they need more to compress for certain benefits to show? (y-axis doesn't go to zero warning)