This paper presents a method to extend matrix-vector multipliers to support the evaluation of non-linear functions. The proposed approach introducesnon-linearity by optionally overriding the input sig-nals of the matrix-vector multiplier. The methodaims to reduce the idleness of hardware resources dur-ing computation, to maximise the reuse of arithmeticunits and internal structures in existing matrix-vectormultipliers, and to reduce the effort in adding addi-tional functions. From our analysis on a case witheight non-linear functions, the proposed design con-sumes fewer components for addition, multiplication,division, exponentiation and logarithm than a refer-ence design with dedicated function evaluation facil-ities.