decimal

D implementation of floating point decimal data type according to IEEE-754-2008

Why

• 1.1 + 2.2 = 3.3 (decimal) but 1.1 + 2.2 = 3.3000000000000003 (float)

• 0.45 = 0.45 (decimal) but 0.45 = 0.450000018 (float)

Features:

• fully IEEE-754-2008 compliant;
• using Intel's binary decimal enconding;
• three decimal data types: decimal32, decimal64 and decimal128
• all D operators supported for all numeric types (left and right side integrals, floats, chars);
• conversion supported from/to integrals, floats, bools, chars
• conversion to other decimal formats (Microsoft Currency, Microsoft Decimal, IBM Densely Packed Decimal)
• all std.math functions implemented (even logarithms and trigonometry);
• all format specifiers implemented (%f, %e, %g, %a);
• integrated with phobos format and conversion functions (to, format, writef);
• thread local precision (from 1 to 34 decimal digits);
• new rounding mode - Europe's most used - tiesToAway;
• alternate exception handling (through flags);
• minimal dependencies (some traits and some floating point functions);
• comprehensive documentation;

Documentation: http://rumbu13.github.io/decimal/doc/decimal.html