On 01/27/2016 12:14 PM, Mike Snitzer wrote: > Yeah, I was using some crap ioping test to get those 500K/85K results. > I've now switched to the fio test Sagi shared [...] Tsk tsk -- Jens Axboe