From: Robert Klemme <shortcutter@...>
Date: 2006-10-19T16:45:13+09:00
Subject: Re: Performance of ruby hashes

On 19.10.2006 09:18, Paul Lutus wrote:
> The basic idea is that hash speed degrades as the size of each single-level
> hash increases, and after a certain size has been reached, breaking the
> hash into groups (layers if you prefer) improves performance.

What is the basis for this statement?  The pure lookup speed certainly 
does not degrade unless there is an increasing number of collisions 
(which should be prevented by the load factor controlling the table's size).

> Give us the details of the problem -- the nature of the keys, the size of
> the data set, and so forth.

Definitively!

Kind regards

	robert