Re: Skybuck presents Faster Div 7

Robert Redelmeier mentioned:

> > But I'm not sure if this code will be any faster than
> > a SUB-LOOP-divider for smaller values or the classical
> > DIV-instruction for values above 16 bit width, even a few
> > minor optimisations could speed it up a bit.  ie: swap the
> > two ADD-lines and dw-align the loop.
>
> 'scuse me for the interruption, but exactly how could this
> iteration be faster than MUL by reciprocals (small LUT) ?

Oh yes, I forgot to mention this option ;)

MUL by biased 2**n/7
****ft right n

or even use an FPU biased MUL/****ft-divide
__
wolfgang