Messages in this thread | | | Date | Mon, 10 Mar 2003 23:47:12 -0800 | From | William Lee Irwin III <> | Subject | Re: Dcache hash distrubition patches |
| |
On Mon, Mar 10, 2003 at 11:41:07PM -0800, Martin J. Bligh wrote: > It's not perfect, but not bad either. Some mathematician can go calculate > just how imperfect it is over random distribution, but it looks OK to me ;-)
Okay, the standards are a bit more stringent for hash functions. You basically want the total number of collisions to be minimized.
That said, standard statistical things are useful for first-pass discrimination.
This looks basically okay, but if it's really just even with respect to cache pressure the need for a different data structure in the eventual future is clear. I'd need a baseline (pre-hlist) dcache snapshot to compare against to get a better idea.
-- wli - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |