Re: [patch 1/13] Qsort

[Jesper Juhl]

On Sun, 23 Jan 2005, Andi Kleen wrote:

> > How about a shell sort?  if the data is mostly sorted shell sort beats 
> > qsort lots of times, and since the data sets are often small in-kernel, 
> > shell sorts O(n^2) behaviour won't harm it too much, shell sort is also 
> > faster if the data is already completely sorted. Shell sort is certainly 
> > not the simplest algorithm around, but I think (without having done any 
> > tests) that it would probably do pretty well for in-kernel use... Then 
> > again, I've known to be wrong :)
> 
> I like shell sort for small data sets too. And I agree it would be 
> appropiate for the kernel.
> 
Even with large data sets that are mostly unsorted shell sorts performance 
is close to qsort, and there's an optimization that gives it O(n^(3/2)) 
runtime (IIRC), and another nice property is that it's iterative so it 
doesn't eat up stack space (as oposed to qsort which is recursive and eats 
stack like ****)...
Yeah, I think shell sort would be good for the kernel.


-- 
Jesper Juhl



-
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/