We describe a universal hash-function family, PolyR, which hashes messages of effectively arbitrary lengths in 3.9-6.9 cycles/byte (cpb) on a Pentium II (achieving a collision probability in the range 2−16-2−50). Unlike most proposals, PolyR actually hashes short messages faster (per byte) than long ones. At the same time, its key is only a few bytes, the output is only a few bytes, and no “preprocessing” is needed to achieve maximal efficiency. Our designs have been strongly influenced by low-level considerations relevant to software speed, and experimental results are given throughout.
CITATION STYLE
Krovetz, T., & Rogaway, P. (2001). Fast universal hashing with small keys and no preprocessing: The polyR construction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2015, pp. 73–89). Springer Verlag. https://doi.org/10.1007/3-540-45247-8_7
Mendeley helps you to discover research relevant for your work.