I think it's hard to compare the results here? In test 1 (single IW shared across threads) you end up with one index.
In test 2 (private IW per thread) you end up with N indexes, which to be "fair" need to be merged down into one index (eg with .addIndexes)? Or seen another way, test 1 should have a much larger merge factor to be comparable to test 2. You should also call IW.close(false) (aborts all running merges) whenever comparing two indexing runs... otherwise a long running merge can completely skew the results. However, there is known contention in IndexWriter's flushing (when it writes the buffered documents to a new segment), which is still present in 3.x but is fixed on trunk (to be 4.0)... see this post: http://blog.mikemccandless.com/2011/05/265-indexing-speedup-with-lucenes.html Mike McCandless http://blog.mikemccandless.com On Tue, Jan 10, 2012 at 8:32 PM, Cheng <zhoucheng2...@gmail.com> wrote: > Hi, > > I use a same instance of writer for multiple threads. It turns out that the > time to finish jobs is more than to create a new writer instance in each > thread. What would be the possible reasons? > > Thanks --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
