[Elecraft] Faster FFT...off topic but relevant

[Tony Estep]

On Tue, Jan 31, 2012 at 2:32 PM, Jessie Oberreuter
<joberreu-elecraft at moselle.com> wrote:
>... it isn't going to change the world...
============
I remember about 30 years ago when the Karmarkar algorithm for linear
optimization appeared. On certain problems, it appeared that it would
be about 10x as fast as conventional schemes. At the place I was
working we were doing lots of big linear problems involving bond
portfolios, and we got all excited. One of the guys laboriously wrote
code to implement it, only to discover that on our kind of problems,
the plain-vanilla simplex approach we had been using was about 10x
faster than the new fancy one.

Tony KT0NY



-- 
http://www.isb.edu/faculty/facultydir.aspx?ddlFaculty=352