DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://issues.apache.org/bugzilla/show_bug.cgi?id=30736>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=30736 [PATCH] to remove synchronized code from TermVectorsReader ------- Additional Comments From [EMAIL PROTECTED] 2004-08-23 20:57 ------- yes i did. The performance improve will be visible having multiple threads fetching term vectors and sharing the same IndexReader. It's hardly to test on a single processor machine and thats all i have for the moment. The tests i did, runs against an index with 25.000 text files (avg. 4KB per file). The test is firing 500 calls per thread to fetch term vectors with a random generated document id. Before i start the real measurement, i'm running several thousand queries against the index to warm up lucene and the file system cache. What you see is that the performance boost is between 0% running a single thread and 30% running 10 threads. I'm sure it would be even better running it on a multiple processor machine. Where i have no explanation for, is that the index creation process is about 10% faster using the patch. best regards Bernhard --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
