Floating-point numbers

TReal is a 64-bit double-precision type, used as the argument type for all math functions, and the default type of floating-point literals.

TReal64 and TReal32 are available when the precise size is important.

EPOC floating-point arithmetic conforms to IEEE754 specification.

Most EPOC implementations do not have hardware floating-point support. As a result, floating-point arithmetic is noticeably less efficient than integer arithmetic. It should be avoided unless demanded by the problem type.