Category Archives: JLinAlg

JLinAlg is an open source and easy to use Java library for linear algebra. It is licensed under the LGPL.

Rounding Errors And Arbitrary Precision Arithmetic

The computer computes. Plus, minus, times and sometimes even divide. The results are always correct. Really? Computers usually use floating point numbers, which are very flexible, but still can cause quite substantial rounding errors – at least when you are doing complex calculations with many steps. These small inaccuracies can accumulate and become a creeping poison. Arbitrary precision arithmetic offers a way out. Continue reading