lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From pfaun <stephan.h...@yahoo.de>
Subject WG: performance boost through multithreaded query processing?
Date Thu, 23 Oct 2008 13:07:03 GMT
Hello,

Currently we are facing the problem that some searches espacially fuzzy (term~0.6) wildcard
searches (*term*) needs some time depending on the field-searchword combination (the more
terms there are the more processing has to be done).
We improved the performance through caching the bitsets of the single fuzzy query/wildcard
query.

Within our logs we can see that combined queries within a BooleanQuery are processed sequentially.
So our question are: Does it make sense for you to parallelize the processing of the queries
within a boolean query (with a restriction of the amount of prallel processed queries)? With
the caches in mind it might be faster and the system is running on a multicore machine. Has
anyone experience in prallelizing single query processing within a BooleanQuery?
Could there be drawbacks combining the results of the booelan clauses (e.g some IO). At the
end there should only be the bitsets connected to the terms and this should be in memory already,
shouldn't it?

Thanks in advance

stephan



      
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message