decimalfloat

Decimalfloat refers to decimal floating-point numbers, a class of numeric representations in which numbers are stored with a base-10 significand and a base-10 exponent. This arrangement allows decimal arithmetic with a fixed number of significant digits while preserving the decimal scale, reducing the rounding errors that can occur when decimal values are represented in binary floating-point. Decimalfloat is designed to give exact decimal results for many common decimal values and to provide predictable rounding behavior for financial and commercial calculations.

Standards and formats commonly associated with decimalfloat come from the IEEE 754 family. The IEEE 754-2008 standard defines decimal floating-point formats such as decimal32, decimal64, and decimal128, which carry 7, 16, and 34 decimal digits, respectively, along with exponent ranges, special values (such as NaN and infinities), and well-defined rounding modes. Implementations may encode the decimal significand in various ways, including binary significands with a decimal exponent or specialized decimal encodings such as densely packed decimal.

Advantages of decimalfloat include precise decimal representation for typical monetary values, exact input and output of decimal fractions like 0.1, and consistent rounding behavior across calculations. These properties make decimalfloat attractive in financial, accounting, and commerce-related software where binary floating-point can introduce subtle errors.

Comparison with other numeric types highlights trade-offs: decimalfloat offers correct decimal arithmetic at the cost of typically higher computational overhead than binary floating-point, and it provides more range and precision control than fixed-point representations. Adoption varies by language and platform, with many ecosystems offering libraries or native support aligned to IEEE 754 decimal formats.