Releases: uofs-simlab/func
Releases · uofs-simlab/func
v2.2.0
What's Changed
- Using Doxygen for generating code documentation
- EqSpaceInterpTable is now ExactInterpTable (hopefully more indicative of what that LUT type does rather than how it's implemented). - A bug with the max table value of nonuniform LUTs has been fixed which will improve their step size vs error curves.
- FailureProofTable uses a slightly better interface (it deduces TIN and TOUT automatically).
Full Changelog: v2.1.0...v2.2.0
v2.1.0
What's Changed
- Dev by @shawn-mcadam in #16
- Composite LookupTable fixes by @shawn-mcadam in #17
- Numerous Bug Fixes and Beta Multivariate LUTs by @shawn-mcadam in #19
Full Changelog: v2.0...v2.1.0
v2.0
What's Changed
- In LookupTableComparator, users can randomly generate inputs to functions based on their own random distributions by @shawn-mcadam in #2
- Recording arguments sampled by a function during a simulation by @shawn-mcadam in #4
- Added a linear algebra library by @shawn-mcadam in #5
- Added Pade approximants by @shawn-mcadam in #6
- Padded polynomials to a power of two by @shawn-mcadam in #7
- Revisions by @shawn-mcadam in #8
- Added autodiff by @shawn-mcadam in #9
- Added composite LUTs by @shawn-mcadam in #10
- Added nonuniform LUTs by @shawn-mcadam in #11
- Added some OpenMP support by @shawn-mcadam in #12
- Templated every class by @shawn-mcadam in #13
- Reworked everything and fixed bugs by @shawn-mcadam in #14
- Reworked everything and fixed bugs by @shawn-mcadam in #15
New Contributors
- @shawn-mcadam made their first contribution in #2
Full Changelog: v1.0...v2.0
v1.0.0
This release is unsupported. This is the version of FunC used in these papers: https://epubs.siam.org/doi/10.1137/18M1201421 and https://www.sciencedirect.com/science/article/pii/S136481522100061X.
This is a small library for building linear, quadratic, and cubic interpolation- or Taylor-based LUTs mapping double to double. The only dependencies are Boost and Quadmath. UniformLookupTableGenerator can build LUTs according to a max tolerance for error, and ImplementationComparator can profile LUTs. See examples/ for usage.