[ https://issues.apache.org/jira/browse/CASSANDRA-7244?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Benedict updated CASSANDRA-7244: -------------------------------- Description: In QueryOptions.Codec.Flag.deserialize() we call Flag.values(), which constructs a copy of the Enum array each time. Since we only use this lookup the enums, and since it happens _often_, this is pretty wasteful. It seems to make a slight (barely perceptible) difference to throughput, but worth fixing either way. (was: In QueryOptions.Codec.Flag.deserialize() we call Flag.values(), which constructs a copy of the Enum array each time. Since we only use this lookup the enums, and since it happens _often_, this is pretty wasteful. It seems to make a few % difference to throughput.) > Don't allocate a Codec.Flag enum value array on every read > ---------------------------------------------------------- > > Key: CASSANDRA-7244 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7244 > Project: Cassandra > Issue Type: Improvement > Components: Core > Reporter: Benedict > Assignee: Benedict > Priority: Minor > Labels: performance > Fix For: 2.1 rc1 > > Attachments: 7244.txt > > > In QueryOptions.Codec.Flag.deserialize() we call Flag.values(), which > constructs a copy of the Enum array each time. Since we only use this lookup > the enums, and since it happens _often_, this is pretty wasteful. It seems to > make a slight (barely perceptible) difference to throughput, but worth fixing > either way. -- This message was sent by Atlassian JIRA (v6.2#6252)