[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-26 Thread Ignacio Vera (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16893763#comment-16893763 ] Ignacio Vera commented on LUCENE-8928: -- I tried to see the effect running on 1D ranges and it is

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Robert Muir (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890224#comment-16890224 ] Robert Muir commented on LUCENE-8928: - I guess I'm saying, I don't see any evidence of speedups for

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Ignacio Vera (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890214#comment-16890214 ] Ignacio Vera commented on LUCENE-8928: -- 1-dimensional ranges (2D in total).  In fact the shapes

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Robert Muir (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890207#comment-16890207 ] Robert Muir commented on LUCENE-8928: - Do you mean 1-dimensional ranges (actually 2D) or

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Ignacio Vera (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890194#comment-16890194 ] Ignacio Vera commented on LUCENE-8928: -- I would expect some increase on QPS for 2D when there is

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Robert Muir (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890186#comment-16890186 ] Robert Muir commented on LUCENE-8928: - Seems like it should only happen when number of dimensions is

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-22 Thread Ignacio Vera (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16890173#comment-16890173 ] Ignacio Vera commented on LUCENE-8928: -- I run this approach locally. It helps as well in the case

[jira] [Commented] (LUCENE-8928) BKDWriter could make splitting decisions based on the actual range of values

2019-07-19 Thread Adrien Grand (JIRA)
[ https://issues.apache.org/jira/browse/LUCENE-8928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16888722#comment-16888722 ] Adrien Grand commented on LUCENE-8928: -- I played with this idea a bit at