Apfloat is an arbitrary precision arithmetic library. Most operations e.g. multiplication can be done with O(n log n) complexity. That is, when the number of digits doubles, the execution time only slightly more than doubles.

Apfloat requires Java 8 or later.

The apfloat library has been designed for high performance, but as it is 100% Java, the achievement is somewhat limited compared to various other existing libraries written in C and assembler. For best performance it's recommended to use the fastest Java Virtual Machine available, for example Oracle's or OpenJDK's server VM.

Apfloat is distributed under the terms of the GNU Lesser General Public License and comes with NO WARRANTIES.

To report bugs and submit questions or comments, please contact the author Mikko Tommila.