Parallel comba multiplication in GF(2163) using homogenous multicore microcontroller

Comba Multiplication is considered as one of the most important multiplication techniques used in public key cryptography computations be it in modular form in RSA or in finite field form in Elliptic Curve Cryptography for example. The efficiency of these Public Key Cryptography implementations depends heavily on the efficiency of the implementation of the multiplication operation. Multicore architectures are becoming increasingly important platforms for modern computation. However cryptography implementations on these platforms is still at its infancy. In this paper, we propose a parallel software implementation of the comba multiplication in GF(2
¹⁶³
) using a homogenous multicore microcontroller. We obtain performance results and compare these to sequential implementation of comba multiplications with and without modular reduction for different word size 8, 16 and 32 bits on single core microcontrollers. Our obtained results outperform most of the published single core modular multiplication implementations and require much fewer cycles. We achieve more than 85% enhancement of the measured time in comparison to a single core implementation.