Blindingly fast! The maximum absolute error (>0.07 radians) is, alas,
much too big for my application.
Here's another answer to my question, based on a 50-year old atan
approximation due to Hastings. It has |error| < 0.005, and (with all
relevant optimizations in gcc) is 3-5x faster than atan2 in spite of
two horrible divisions in the main code path.
I wrote my own AltiVec function which computes normalized phase
(atan2/pi) and did determine my own coefficients for Ar3-Br-C, which
results in a maximum error of 0.1539% (0.00483 rads), so I guess that's
at least as good as any of the other fast estimate methods... ;)
--
Shaun Wexler
MacFOH http://www.macfoh.com
