Re: random i/o speedup weirdness

On Mon, May 14 2007, Andras.Horvath_at_cern.ch wrote:
> > Probably because fio gives up finding new free random blocks and then it
> > sort-of becomes a bit more sequential. Can you check if adding
> > 'norandommap' to the options makes that effect go away?
>
> just finished testing -- yes, 'norandommap' makes the effect go away
> while producing consistent results with the original test.

Great, I think I'll try making that the default option then, or at least
fixup the randommap behaviour. Using a random map is mainly beneficial
for write verifying.

-- 
Jens Axboe