> Micha=B3 Pasternak wrote:
> > > > > performance."
> > > > >
> > > > > The NetBSD approach is "Detect the controller and HD, and automat=
ically
> > > > > configure for the highest throughput supported by both devices."
> > > >
> > > > Interesting, how come Linux beats NetBSD at the HD performance on i=
386
> > > > in _all_ cases?
>
> Would be interesting to see the difference - would you share the
> benchmarking details? I suppose you use ext2fs on Linux
> and async mounted ffs on NetBSD to get fair comparison.
>
NetBSD wins in all cases i've seen - highest difference is on slowest
machines.