* Arne H Juul <Arne.Juul@osl.fast.no> [011211 17:55] wrote:
> On Tue, 11 Dec 2001, Ignatios Souvatzis wrote:
>  > On Mon, Dec 10, 2001 at 07:42:07PM +0100, Wojciech Puchar wrote:
>  > > > this is more often called "direct i/o" or "unbuffered i/o".
>  > > > it's on the list of things I want to add, but I haven't had time yet.
>  > > 
>  > > it would be really good! with support in common mpeg/divx players etc..
>  > > and with option in cp and dd
>  > 
>  > Why do you think that "raw" I/O would be more efficient than the filesystem
>  > + buffer cache doing readahead? 
> 
> Because that's what benchmarks and real production use of some
> applications with and without direct io show :-)  I don't think
> "cp" is such a good example though, since you lose the readahead.
> For some (typically database-like) apps it can be a pretty big boost.
> 
> Disclaimer:  We've actually *measured* the performance difference
> for our app only on FreeBSD, Windows, Solaris, and Linux, AFAIK.

Direct IO is a good idea, it prevents applications that are basically
doing a cache walk from blowing out all your cache.

Take for instance anything that either keeps its own cache (postgresql)
or something that would completely never reuse its cache, like
cp/dd.

If you want to check out performance of your app on FreeBSD give the
open(2) syscall a flag of O_DIRECT and let me know how it changes
your numbers. :)

-- 
-Alfred Perlstein [alfred@freebsd.org]
'Instead of asking why a piece of software is using "1970s technology,"
 start asking why software is ignoring 30 years of accumulated wisdom.'
                           http://www.morons.org/rants/gpl-harmful.php3