Hash index is well used index in equality queries. But it perform no better than B-tree in PostgreSQL.
  
  The PostgreSQL community has discuss the problem before

  http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php

  This project tries to deal with this problem by storing the hash value instead of real key. It's one of schemes to improve hash index performance listed in PostgreSQL's TODO list.

  This work will deal with this problem by storing the hash value instead of the real key. Then I proposed an idea to deal with the dark side of the scheme. I'll benchmark both of them against original implementation and B-Tree to show the improvement.