Estudo e otimização do LowMC

Abstract

The evolution of computing is always accompanied by new challenges and one of these are the problems considered intractable for classical computing that become treatable in quantum computing. One of these problems directly affects a subarea of computing, which is cryptography; fundamental in the secure communication of digital devices. In 1994, Peter Shor presented a quantum algorithm capable of solving the problem of factoring integers and the discrete logarithm. This discovery was a milestone to determine the beginning of studies in the area of post-quantum cryptography. Such problems, arising from number theory, are the basis for asymmetric cryptographic primitives. Thus, a digital signature algorithm emerged based on the concept of Proof of Knowledge and Multipart Computing called Picnic, such an algorithm is considered secure even with the emergence of a general-purpose quantum computer, but this algorithm is still relatively slow. Within Picnic there is a symmetric cipher that runs a few hundred times independently. Thus, this work aims to seek optimizations in this specific cryptographic primitive, seeking to contribute to cryptography in general. The main technique applied is instruction-level parallelization, to use instruction vectors, to process on amounts of bits in parallel.