Re: [PATCH]: Free pages from local pcp lists under tight memoryconditions
From: Christoph Lameter
Date: Wed Nov 23 2005 - 17:01:46 EST
On Wed, 23 Nov 2005, Andrew Morton wrote:
> I was able to demonstrate a large (~60%?) speedup in one microbenckmark
> which consisted of four processes writing 16k to a file and truncating it
> back to zero again. That gain came from the cache warmth effect, which is
> the other benefit which these cpu-local pages are supposed to provide.
Maybe we can cut the pcp logic back to only put cache warm pages into a
single per_cpu pcp list for the local processor that contains node local
pages? Return all remote pages and cold pages directly to the buddy lists.
That way we can remove the logic to flush remote pages and remove those
pcp lists for remote nodes that are mostly empty anyways.
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/