Arbitrary-precision dyadic numbers and interval arithmetic, via the types:

  • Dyadic: variable-precision floats with exact ring operations

  • MPBall: float ± error bound with field & elementary (interval-like) operations

By default, using hmpfr when compiling with ghc >=7.10. Older versions of ghc are no longer tested.

Set flag MPFRRounded to bind MPFR not via hmpfr but via a tweaked version of package rounded:

This backend is considerably faster than hmpfr. It is likely to become the default in future.

There is a plan to add an Integer-only backend so that this library can be used without MPFR.


  • v 2017-11-14 * fix compilation with haskell-mpfr
  • v 2017-11-14 * using Claude Heiland-Allen's Numeric.Rounded.Simple
  • v 2017-09-12 first release on Hackage backends: hmpfr and (tweaked) rounded
comments powered byDisqus