Abstract: A method of modular reduction P mod Q based on a precomputation. Two quantities are precomputed and stored in a look-up table. Based on these stored precomputed quantities, two stages of partial modular reduction are performed. The result of the second stage of partial modular reduction is then adjusted according to the value of Q, to determine the modular reduction result. The modular reduction result may be provided to variable radix multiplication logic circuitry.