[
https://issues.apache.org/jira/browse/LUCENE-3003?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13013869#comment-13013869
]
Dawid Weiss commented on LUCENE-3003:
-------------------------------------
For what it's worth, the instrumentation interface allows one to get exact
allocation sizes of objects. I put together a small spike at
https://github.com/dweiss/poligon/tree/master/instrumenter that measures the
actual allocation size of byte[]. On my hotspot, 64-bit, this yields:
{noformat}
byte[0] takes 24 bytes.
byte[1] takes 32 bytes.
byte[2] takes 32 bytes.
byte[3] takes 32 bytes.
byte[4] takes 32 bytes.
byte[5] takes 32 bytes.
byte[6] takes 32 bytes.
byte[7] takes 32 bytes.
byte[8] takes 32 bytes.
byte[9] takes 40 bytes.
byte[10] takes 40 bytes.
byte[11] takes 40 bytes.
...
{noformat}
IBM's VM yields the same (64-bit), but the version of jrockit that I have
(which may be an old one, but is 64-bit!) yields:
{noformat}
byte[0] takes 16 bytes.
byte[1] takes 24 bytes.
byte[2] takes 24 bytes.
byte[3] takes 24 bytes.
byte[4] takes 24 bytes.
byte[5] takes 24 bytes.
byte[6] takes 24 bytes.
byte[7] takes 24 bytes.
byte[8] takes 24 bytes.
byte[9] takes 32 bytes.
byte[10] takes 32 bytes.
byte[11] takes 32 bytes.
byte[12] takes 32 bytes.
byte[13] takes 32 bytes.
byte[14] takes 32 bytes.
byte[15] takes 32 bytes.
byte[16] takes 32 bytes.
byte[17] takes 40 bytes.
{noformat}
Don't have access to a 32-bit system right now, but if you're keen on checking,
checkout that github repo and run:
{noformat}
cd instrumenter
mvn package
java -javaagent:target/instrumenter-0.1.0-SNAPSHOT.jar -version
{noformat}
> Move UnInvertedField into Lucene core
> -------------------------------------
>
> Key: LUCENE-3003
> URL: https://issues.apache.org/jira/browse/LUCENE-3003
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Reporter: Michael McCandless
> Assignee: Michael McCandless
> Priority: Minor
> Fix For: 3.2, 4.0
>
> Attachments: LUCENE-3003.patch, LUCENE-3003.patch
>
>
> Solr's UnInvertedField lets you quickly lookup all terms ords for a
> given doc/field.
> Like, FieldCache, it inverts the index to produce this, and creates a
> RAM-resident data structure holding the bits; but, unlike FieldCache,
> it can handle multiple values per doc, and, it does not hold the term
> bytes in RAM. Rather, it holds only term ords, and then uses
> TermsEnum to resolve ord -> term.
> This is great eg for faceting, where you want to use int ords for all
> of your counting, and then only at the end you need to resolve the
> "top N" ords to their text.
> I think this is a useful core functionality, and we should move most
> of it into Lucene's core. It's a good complement to FieldCache. For
> this first baby step, I just move it into core and refactor Solr's
> usage of it.
> After this, as separate issues, I think there are some things we could
> explore/improve:
> * The first-pass that allocates lots of tiny byte[] looks like it
> could be inefficient. Maybe we could use the byte slices from the
> indexer for this...
> * We can improve the RAM efficiency of the TermIndex: if the codec
> supports ords, and we are operating on one segment, we should just
> use it. If not, we can use a more RAM-efficient data structure,
> eg an FST mapping to the ord.
> * We may be able to improve on the main byte[] representation by
> using packed ints instead of delta-vInt?
> * Eventually we should fold this ability into docvalues, ie we'd
> write the byte[] image at indexing time, and then loading would be
> fast, instead of uninverting
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
