PP NUMATH - A program package for methods of numerical mathematics. With remarks on matrix processor variants A general description of the code PP NUMATH, developed at the Karl Weierstrass Institute of Mathematics of GDR Academy of Sciences is given. This is a general purpose code for solving problems in numerical mathematics, e.g., linear algebra, including systems of linear equations, linear last square problems, eigenproblems; approximation, interpolation and numerical differentiation of functions, based on polynomials, splines and trigonometric polynomials; numerical integration, including multiple integrals, solving nonlinear equations; minimization of functions of one and several variables; linear, quadratic and nonlinear optimization; ordinary differential equations, integral equations and partial differential equations. The code has a hierarchical structure: problem oriented solvers, method oriented primary routines and basic modulus for matrix and vector handling. By changing, for example, the basic modulus one can adapt easily the code to vector or array processor. Some numerical tests for the speedup reached on a vector processor for this purpose oriented matrix and vector handling modulus are demonstrated.