Re: speed of octave symbol table code

[John W. Eaton]

On 23-Oct-2007, Tom Holroyd wrote:

| On Tue, 2007-10-23 at 00:39 -0400, John W. Eaton wrote:
| > Increasing
| > the table size by 50% gives us more free chains, but the hash function
| > doesn't seem to be spreading the symbols out to fill the additional
| > chains and the maximum chain length (which affects lookup speed) ends
| > up being about the same.
| 
| Have you tried the FNV hash? It's pretty simple and works well.
| 
| http://www.isthe.com/chongo/tech/comp/fnv/

No, but I don't think it is worth changing this now.  We are about to
release 3.0, and the hashing code will be replaced by something else
in 3.1.

jwe