Re: 2.6.4-mm1

[Andrew Morton]

jlnance@xxxxxxxxxxxxxx wrote:
>
> On Wed, Mar 10, 2004 at 11:31:40PM -0800, Andrew Morton wrote:
> >   This affects I/O scheduling potentially quite significantly.  It is no
> >   longer the case that the kernel will submit pages for I/O in the order in
> >   which the application dirtied them.  We instead submit them in file-offset
> >   order all the time.
> 
> Hi Andrew,
>     I have a feeling this change might significantly improve the external
> sorting benchmark I emailed you ( http://lkml.org/lkml/2003/12/20/46 ).
> I will try running it when I get a chance and let you know.

That thing's still sitting in by Inbox awaiting attention :(

I just took a quick peek.  The `testfile' file which it lays out is well
laid-out so yes, if you're seeking all over that file then 2.6.4-mm1 may
indeed help throughput.

But `tesfile.tmp' is not well laid-out.  Looks like it was created seekily,
so its blocks are all jumbled up.

The code in 2.6.4-mm1 favours unjumbled-up files.  The code in 2.4 and
2.6.4 favours jumbled-up files.

When kjournald performs writeback it favours jumbled-up files, even in
2.6.4-mm1.

So it's hard to say what will happen ;)  I'll take a look later.

> It gives me a good excuse to get 2.6 kernels working on my systems :-)

Luddite.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/