Because of the recent good results for pgbench on DragonFly 3.2, Phoronix has a new benchmark of DragonFly using other (possibly unrelated) tests.  There’s not a lot of information to glean from them; they are testing operations different than what was optimized for pgbench in 3.2.  I’d like to see DragonFly 3.0 tested the same way to see how much improvement there was between versions.