Item — HN Reader

by pella·1y ago·view on hn ↗

Probably, the results in the table are about 7 years old. For example, the table shows version 1.3.4 for zstd, which indicates that the results are outdated - currently, version 1.5.6 is the latest release ( https://github.com/facebook/zstd/releases/tag/v1.5.6 )

In my opinion, it is better to check the original repository: https://github.com/inikep/lzbench

or this: https://github.com/inikep/lzbench/blob/master/doc/lzbench20_...

Judging by the last link, the only options that are strictly superior to snappy in every sense are lz4 (as OP claimed) and zstd (levels -3 to +1). If compression speed is critical, then the default level of zstd compression in most big data tools (+3) might be the wrong one for you, but test on your own data, not Silesia.