next up previous
Next: Target precision Up: Variable precisions Previous: Rational

Arbitrary-precision floats

Variable (high) precision arithmetic: CAS have these, but you may also just use Fortran or C subroutines: MPFUN (D. H. Bailey) is only the most recent of a sequence of library packages to support

Designs like MPFUN tend to present (at least in their design) new explicit choices that were more easily settled for fixed precisions.

E.g. SIN( argument, precision-of-result, value-of-pi) where value of pi is needed for argument reduction.

(Fortran also may force other silly global choices ``max precision for this run of MPFUN'' compiled into COMMON)



Richard J. Fateman
Thu Aug 13 21:54:23 PDT 1998