Re: 2.6.13-mm3

[Andrew Morton <akpm@osdl.org>]

Sonny Rao <sonny@burdell.org> wrote:
>
> On Mon, Sep 12, 2005 at 02:43:50AM -0700, Andrew Morton wrote:
> <snip>
> > - There are several performance tuning patches here which need careful
> >   attention and testing.  (Does anyone do performance testing any more?)
> <snip>
> > 
> >   - The size of the page allocator per-cpu magazines has been increased
> > 
> >   - The page allocator has been changed to use higher-order allocations
> >     when batch-loading the per-cpu magazines.  This is intended to give
> >     improved cache colouring effects however it might have the downside of
> >     causing extra page allocator fragmentation.
> > 
> >   - The page allocator's per-cpu magazines have had their lower threshold
> >     set to zero.  And we can't remember why it ever had a lower threshold.
> > 
> 
> What would you like? The usual suspects:  SDET, dbench, kernbench ?
> 

That would be a good start, thanks.  The higher-order-allocations thing is
mainly targeted at big-iron numerical computing I believe.

I've already had one report of fragmentation-derived page allocator
failures (http://bugzilla.kernel.org/show_bug.cgi?id=5229).