作者: Daniel Lemire , Owen Kaser
DOI: 10.1007/S13389-015-0110-5
关键词:
摘要: Intel and AMD support the carry-less multiplication (CLMUL) instruction set in their x64 processors. We use CLMUL to implement an almost universal 64-bit hash family (CLHASH). compare this new with what might be fastest on processors (VHASH). find that CLHASH is at least 60 % faster. also a popular function designed for speed (Google’s CityHash). 40 % faster than CityHash inputs larger 64 bytes just as fast otherwise.