Vectorization of Poly1305 Message Authentication Code

This paper describes software optimizations for the message authentication code Poly1305. We leverage the wide vectorization capabilities of the new AVX2 architecture, to speed up the authenticator generation on the latest x86_64 processors. In addition, we show how to apply vectorization for the future AVX-512 architecture to achieve a further speedup. The proposed optimization leads to significant performance gains. For example, on the latest Intel Has well micro architecture, our AVX2 implementation performs at 0.66 cycles per byte (on a 4KiB message), which is 2x faster than the current implementation in Open SSL.