Re: Perfect hash function generator

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Wed, Jul 15, 2009 at 01:58:13PM +0100, Mindaugas Rasiukevicius wrote:
> Joerg Sonnenberger <joerg%britannica.bec.de@localhost> wrote:
> > It is not the fastest hash if used standalone. It does ensure that you
> > can get up to three different results with a single pass over the data,
> > which makes compete with MurmurHash2.
> 
> Do you have any numbers?

On AMD64, which is known to have a very fast multiplication:
jenkins_vector_hash: 8.66s
MurmurHash2_64: 5.71s

For the applications, two calls for MurmurHash2_64 are needed.

> > I think the tools used to generate functions in base should be in base
> > as well.
> 
> It is quite specific tool, not really a general purpose functionality.
> Sure it is small, but just breaks the principle to keep base minimal.

It addresses a specific problem domain. It is meant to be used for
programs in base. Requiring 3rd party software when modifying source is
generally be frowned upon.

Joerg