> Hi all !
>
> Yesterday we tested some new srqt functions. First I tried a quick
> hack after the suggested newton-approximation algorithm and it was three
> times as fast as the builtin function.
>
care should be taken when creating functions to calculate values like
SQRT, SIN, COS etc. problems can arise with the distribution and
magnitude of errors.
A number of years(I will be realy dating myself here) AMD produced a
Floting Point co-processor chip. in the documentation for that chip there
was a quite good treatment of the algorithems used and the level of
errors that could be expected. This may be a good starting place for
creating a floating point library that is both reasonably fast and accurate.
Alvin Starr || voice: (416)493-3325
Interlink Connectivity || fax: (416)493-7974
alvin@iplink.net ||