[ https://issues.apache.org/jira/browse/SPARK-1321?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Reynold Xin resolved SPARK-1321. -------------------------------- Resolution: Fixed > Use Guava's top k implementation rather than our custom priority queue > ---------------------------------------------------------------------- > > Key: SPARK-1321 > URL: https://issues.apache.org/jira/browse/SPARK-1321 > Project: Apache Spark > Issue Type: Improvement > Reporter: Reynold Xin > Assignee: Reynold Xin > Priority: Critical > Fix For: 1.0.0 > > > Guava's top k implementation (in Ordering) is much faster than the > BoundedPriorityQueue implementation for roughly sorted input (10 - 20X > faster), and still faster for purely random input (2 - 5X). > We should switch to using that in Spark. -- This message was sent by Atlassian JIRA (v6.2#6252)