> > Well, they could have made xfs go faster with some mkfs and mount options,
> > this is usually what happens in benchmarks which just test out of the box
> > performance.
> Do you mean:
> * Well, they were capable of making xfs go faster with some mkfs and
> mount option but did not
Yes, if they had used different mkfs options and mount options than the
default then they could have made their tests go faster. I am not
implying that they did or did not know they existed.
The reason these options are not the default is that unless you are going
to be pounding on the filesystem in a metadata intensive manner they will
have a negative effect. A bigger log means slower mount and recovery, more
incore log buffers means more memory chewed up.
> * they probably did make xfs go faster with some mkfs and mount options
> Unfortunately could have is very ambiguous :-(
> "The greatest thing you'll ever learn is
> just to love and to be loved in return."
> - David Bowie (Nature Boy from Moulin Rouge)