This package provides a form of compensated arithmetic that lets you perform many operations on chains of doubles at higher precision.
In particular addition, multiplication, etc. on a
Compensated Double is done with twice as many bits of
significand as on a raw
This construction is designed so that it can be iterated.
This was forked from the analytics repository after repeated requests.
Work still needs to be done to properly support most
Floating operations, but everything below that level of the numeric hierarchy should work.
Contributions and bug reports are welcome!
Please feel free to contact me through github or on the #haskell IRC channel on irc.freenode.net.
- Add a benchmark suite.
- Support building with
log-domain-0.13by removing the
log-domain-0.13has removed the
Preciseclass since its methods are now available in
base-4.9/GHC 8.0. As a result, this release drops support for earlier versions of
- Allow building with
- Add a library dependency on the
cabal-doctest. This makes it build with
Cabal-2.0, and makes the
doctests work with
cabal new-buildand sandboxes.
- Support for newer
- Updated for
- Bug fix in
- Marked this package
- Major version bump due to large semantics change in
generic-derivingversion bounds for GHC HEAD compatibility.
- Added a
- Fixed a bug in the
- Reformatted haddocks
- Split from analytics