> What I want to see is the total IO + CPU time across libraries for my specific IO and CPU constraints.
This knob is typically called as a (compression) level, even though it isn't advertised as such because it is hard to translate level into your metric. Some libraries like zstd support IO-adaptive operations as a simpler alternative too.
This knob is typically called as a (compression) level, even though it isn't advertised as such because it is hard to translate level into your metric. Some libraries like zstd support IO-adaptive operations as a simpler alternative too.