Martin Wiesner created OPENNLP-1759:
---------------------------------------
Summary: Optimize computation of hashCode in StringList
Key: OPENNLP-1759
URL: https://issues.apache.org/jira/browse/OPENNLP-1759
Project: OpenNLP
Issue Type: Improvement
Affects Versions: 2.5.4
Reporter: Martin Wiesner
Assignee: Martin Wiesner
Fix For: 2.5.5, 3.0.0
Right now, StringList#hashCode() computes the hash code in a CPU-intensive way
each time it is called. When used in NgramModel and Dictionary, this causes
many extra CPU cycles. Those can be easily avoided, by simply pre-computing the
hashCode for each StringList object once, that is during objection creation as
last step.
Let's tune this towards being more eco-friendly.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
