From @nemequ on March 31, 2015 17:29

It should be possible to make the dataset the variable we compare (i.e., choose a codec and machine, it tells you how the that codec and machine behave on different datasets).

Copied from original issue: quixdb/squash-benchmark#10