Age | Commit message (Collapse) | Author | |
---|---|---|---|
2014-10-07 | Add test code to check reliablity of fixexp(), for sufficently large absolute values (>0.00013) our error is below 0.2%)HEADmaster | erdgeist | |
2014-10-07 | Adding test code for fixexp. We're accurate to 22 significant bits | erdgeist | |
2014-10-06 | Replace rest of hex constants with Q26 macros, also replace algorithm for arguments of fixexp with negative parameter with a more aggressive one | erdgeist | |
2014-10-06 | Add code to look for an optimal solution for negative parameters to exp() ... turns out, the naive solution had the same score as all the other optima. | erdgeist | |
2014-10-06 | Convert hex constants to double constants and a macro to convert them to Q26 | erdgeist | |
2014-10-06 | Clean up table generator | erdgeist | |
2014-10-05 | Fix typo in license comment | erdgeist | |
2014-10-05 | Initial version. Works. Untested. Unreviewed | erdgeist | |