Résumé
This article presents dedicated hardware arithmetic operators for function evaluation. The proposed solution uses polynomial approximations with sparse coefficients which leads to efficient hardware implementations. Up to x2 faster and x8 faster operators are reported compared to standard implementations.