Skip to main content Skip to docs navigation
View on GitHub

Fixed Point

Package for dealing with Big Numbers both integral and decimal

Document reviewed for 5.3.0
On this page

Fixed Point Arithmetic

The Float and Double types serve for handling numbers with decimals. Though its variable precision (there is more density in representable numbers in the 0-1 range, and then quickly decreases) sometime lead to inconsistencies or problems. So for simulations for example, the same parameters and operations might lead to different values in an Intel, AMD, ARM device or with SIMD optimizations.

Fixed Point arithmetic circumvents that problem by implementing deterministically arithmetic with plain integer arithmetic, representing some bits as the integral part and others for the decimal part.

Fixed, FixedLong and FixedShort

Fixed in decimal it has precision of:* +-9999999.99 integer digits, 2 decimal digits. 1 bit for sign, 23~24 bits of integer, 6~7 bits of decimal

inline val Number.fixed: Fixed
val String.fixed: Fixed

FixedLong in decimal it has precision of:* +-99999999999999.9999, 14 integer digits, 4 decimal digits. 1 bit for sign, 47~48 bits of integer, 14~15 bits of decimal

inline val Number.fixedLong: FixedLong
val String.fixedLong: FixedLong

FixedShort in decimal it has precision of:* +-3275.9, 3.3 integer digits, 1 decimal digit. 1 bit for sign, 11~12 bits of integer, 3~5 bits of decimal

fun String.toFixedShort(): FixedShort
inline fun Number.toFixedShort(): FixedShort
Was this article useful?