Re: [PATCH] Radix-tree pagecache for 2.5

[Momchil Velikov <velco@fadata.bg>]

>>>>> "Anton" == Anton Blanchard <anton@samba.org> writes:

Anton> Hi Ingo,

>> Yes, it's very nice. Anton Blanchard has benchmarked both patch variants
>> (tree vs. scalable-hash page buckets) for SMP scalability against the
>> stock hash, on big RAM, many CPUs boxes, via dbench load. He has found
>> performance of radix trees vs. scalable hash to be at least equivalent. (i
>> think Anton has a few links to show the resulting graphs.)

Anton> Here are some results on a 12 way machine. (2.4.16-splay is the radix
Anton> patch):

Anton> http://samba.org/~anton/linux/pagecache_locking/1/summary.png

A correction, "-splay" is the very first variant I posted, which used
splay trees for the page cache.

Anton> As you can see both patches give pretty much equal improvements.

Anton> The other problem with the current pagecache hash is that it maxes out
Anton> at order 9 (due to the get_free_pages limitation) which starts to hurt
Anton> at 4GB RAM and above. On a 32GB machine the average hashchain depth
Anton> was very high:

Anton> http://samba.org/~anton/linux/pagecache/pagecache_before.png

Anton> There were a few solutions (from davem and ingo) to allocate a larger
Anton> hash but with the radix patch we no longer have to worry about this.

Anton> So the radix patch solves 2 problems quite nicely :)

Anton> Anton