Abstract: The main work in association rule mining is how to find mining frequent itemsets efficiently．When the exis-ting frequent itemsets mining algorithms are analyzed,it can be concluded that the computing scale will increase sharply with the increase of data scale．So association rule mining in large data sets on PC platform is difficult to get desired result．Combining the advantages of Eclat and dEclat algorithm,PEclat,an efficient parallel frequent itemset mining algorithm was presented．PEclat runs in the multi-core and shared-memory computing environment．It works as task-level parallel and can process large data sets．The experiment states that better performance can be achieved regardless of the data density.

Key words: Chip multi-core,Frequent itemsets,Parallel process,Association rule