On Tue, 19 Jun 2001, Robert F Schaefer wrote:

> ISTR a thread on c.o.v that indicated a 10x speed improvement for a fella
> who rolled his own client w/ native VAX vector math.  Of course, since it
> wasn't an `official' client, the work was rejected.  I can't recall if it
> was project bovine, or SETI.

Yeah, but this would also mean that you'd have to have one of the more
high-end systems which supported the VAX vector processors.

-brian.