sysbench / fileio - Linux vs. FreeBSD

[Attilio Rao]

2010/6/5 Matthew Jacob <mj at feral.com>
>
> All of these tests have been apples vs. oranges for years.
>
> The following seems to be true, though:
>
> a) FreeBSD sequential write performance in UFS has always been less than optimal.
>
> b) Linux sequential write performance in just about any filesystem has always been "impressive". But that "impressive" has come at some not so obvious costs. First of all, Linux is probably the most aggressive cluster/write-behind OS I've even seen. You can suck down all available memory with writebehind using dd. This means that some stats are "impressive", and others are "painful". A desktop that becomes completely unresponsive while you're doing this dd is one personal outcome.
>
> Also, you have to be careful what you're asking for in comparing the two platforms, or any platforms for that matter. What do you want to optimize for? Apparent responsiveness as a desktop? A specific workload (nfs, cifs) that completes N quatloos per fortnight?

Besides anything, I'm much more concerned about the loss of
performance within FreeBSD itself. I wouldn't expect a so high
pessimization when the number of threads increases (without
considering the big performance loss with the 8k blocksize, pretty
much reproducible). I'm trying to drive, privately, the tester to
pmc/lock profiling analysis in order to start collecting some useful
datas.
While I think that we might pay a lot of attention to ZFS, I think we
might not leave alone FFS. Having a fast, well supported, native
filesystem might be a great thing for us.

Comparing with other operating systems, as you smartly point out,
might not be got as 'undefeatable truths' but have cons and prons that
needs to be fully understood before to make false claims.

Thanks,
Attilio


--
Peace can only be achieved by understanding - A. Einstein