ABSTRACT
This chapter presents an assortment of numerical routines and primitives often required in floating-point packages, numerical applications, graphics, and high-level languages. These include the so-called calculator operations, solution of quadratic equations, operations on complex numbers, and conversions between polar and Cartesian coordinates. Many scientific and engineering problems require the evaluation of hyperbolic functions. These functions, called the hyperbolic tangent, hyperbolic sine, and hyperbolic cosine, have similar properties as the circular trigonometric functions, except that the hyperbolic functions are related to the hyperbola rather than to the circle. Software often needs to determine which of two data items is the larger one. In one version of this operation the data items are ordered in the math unit stack. The Intel math units have no provisions for handling imaginary or complex numbers. Complex arithmetic must be implemented in software. Arithmetic operations of addition, subtraction, multiplication, division, powers, roots, and absolute value have been defined for complex numbers.
