Age | Commit message (Collapse) | Author | |
---|---|---|---|
2023-01-06 | pl/math: Update copyright years | Joe Ramsay | |
All files in pl/math updated to 2023. | |||
2022-04-19 | pl/math: Add Vector/Neon log10f | Pierre Blanchard | |
- Neon log10f uses the same approach as math/v_logf but uses coefficients associated with a polynomial approximation of log10(1+x), see pl/math/tools/v_log10f.sollya to reproduce coefficients. - Extended precision can be used to get a 1ulp variant, incurring a performance penalty. - The maximum measured ULP error is 3.31ulps. - A sollya file for scalar log10f is also provided. - Copy math/v_math.h into pl/math/ in case we need to specialize behavior. |