[ 
https://issues.apache.org/jira/browse/LUCENE-4609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13559833#comment-13559833
 ] 

Shai Erera commented on LUCENE-4609:
------------------------------------

Perhaps this should belong to a different issue? I mean, this one is focused on 
a PackedInts encoder/decoder (or any other encoder/decoder that's better than 
VInt).

Separately, it's interesting that this performs worse than DV Source + 
decoding. I mean, the results don't factor in reading and populating the cache, 
right? The test is already "hot" when it's measured?

I must say that I'm not entirely surprised ... having recently looked at 
PackedInts API, it doesn't look so optimized (working w/ DataInput for 
example), where the dgap+vint that we have in CountingFC is *very* optimized. I 
think that what we need is a custom cache, which encodes things similar to 
PackedInts ... or maybe as you say, the bulk read methods would do better.

But we should explore that on another issue?
                
> Write a PackedIntsEncoder/Decoder for facets
> --------------------------------------------
>
>                 Key: LUCENE-4609
>                 URL: https://issues.apache.org/jira/browse/LUCENE-4609
>             Project: Lucene - Core
>          Issue Type: New Feature
>          Components: modules/facet
>            Reporter: Shai Erera
>            Priority: Minor
>         Attachments: LUCENE-4609.patch, LUCENE-4609.patch, LUCENE-4609.patch
>
>
> Today the facets API lets you write IntEncoder/Decoder to encode/decode the 
> category ordinals. We have several such encoders, including VInt (default), 
> and block encoders.
> It would be interesting to implement and benchmark a 
> PackedIntsEncoder/Decoder, with potentially two variants: (1) receives 
> bitsPerValue up front, when you e.g. know that you have a small taxonomy and 
> the max value you can see and (2) one that decides for each doc on the 
> optimal bitsPerValue, writes it as a header in the byte[] or something.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to