From: "Mauricio Fernández" <batsman.geo@...>
Date: 2002-08-29T08:35:13+09:00
Subject: Re: ANN: RJudy-0.1 - Judy Arrays for Ruby

On Wed, Aug 28, 2002 at 09:04:01PM +0900, Christoph wrote:
> "Mauricio Fern�ndez"  wrote in,
> ....
> > I later found some bugs in my plug-in replacement for Hash. I have
> > another patch which apparently fixes everything. It passes the same
> > tests JudyHash does and some others, as st_* is used for everything and
> > Ruby seems to run fine now.
> 
> Hi,
> 
> I am very curious (and running windows so no judy yet) how this
> replacement effect pure method lookup time?  Keeping it
> simple, what are the Fibonacci and Ackerman benchmarks?
[...]

Some more thoughts about this:
 * fib and ack aren't really testing method lookup, as is it cached the
   first time
 * Judy is better as the data structure size grows... method tables are
   small, so it will hardly make any difference against the regular hash
 * in fact these tables are so small the plain hash table wouldn't have
   to be re-allocated and rehashed to grow any further...
 * JudySL-style hashes work only when (key1 == key2) <=> bit by bit
   identical 

This particular thing (hash indexed by ID, which is a 32 bit (64 where
supported?) integer) should be done with JudyL. If there's any place
using Judy would immediately pay off, it's here. I might work on this
later.

-- 
 _           _                             
| |__   __ _| |_ ___ _ __ ___   __ _ _ __  
| '_ \ / _` | __/ __| '_ ` _ \ / _` | '_ \ 
| |_) | (_| | |_\__ \ | | | | | (_| | | | |
|_.__/ \__,_|\__|___/_| |_| |_|\__,_|_| |_|
	Running Debian GNU/Linux Sid (unstable)
batsman dot geo at yahoo dot com
  
Being overloaded is the sign of a true Debian maintainer.
	-- JHM on #Debian