All of lore.kernel.org
 help / color / mirror / Atom feed
* Fio 1.52
@ 2011-04-29  9:44 Jens Axboe
  0 siblings, 0 replies; only message in thread
From: Jens Axboe @ 2011-04-29  9:44 UTC (permalink / raw)
  To: fio@vger.kernel.org

Hi,

I have tagged version 1.52 of fio. There's not much exciting since
1.51, but one change in particular warrants a note.

As of 1.52, the random generator used for generating IO offset
for random workloads has been changed. We used to rely on
the one supplied by glibc (and similar on other platforms),
but this has a few down sides:

- We did not necessarily generate the same sequences on
  all platforms.
- Our internal generator is faster and from some testing
  of higher quality than the glibc provided one.

So on the plus side we now generate the same IO patterns
on all platforms. The only real downside is that we do not
generate the same IO patterns as 1.51 and previous. This
_should_ not cause any issues, since the pattern should be
equally random. But it will be different. So it is something
to keep in mind, if people use fio for regression runs and
watch performance. We are not at the point where manufacturers
will optimize for a fio workload, so I don't envision any
issues from this change :-)

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2011-04-29  9:44 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-04-29  9:44 Fio 1.52 Jens Axboe

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.