I would also like to see a "control" test run. I'm curious to know if it's not anything specifically in lucene, but rather the performance of the garbage collectors. There are a lot of objects created and destroyed in building a index.
On Thu, May 14, 2009 at 9:05 AM, Granroth, Neal V. <[email protected]> wrote: > It would also be interesting to compare Java vs. Mono performance on Linux. > > When constructing the index it might be good to have several different > document sources. Performance of an index with a relatively small number of > shared terms might be significantly different than an index with a very large > number of shared terms. The presence of "stored" but not indexed document > fields should also be considered. > > When querying the index, don't forget numeric range clauses and query filters. > > When designing the test code it might also be helpful to have a way to > distinguish between general Java vs. .NET performance differences and the > performance differences due specifically to Lucene. > > > -- Neal > > > -----Original Message----- > From: Shashi Kant [mailto:[email protected]] > Sent: Wednesday, May 13, 2009 7:53 PM > To: [email protected] > Subject: Re: Lucene Java Vs .Net > > I am running all tests on the same box - a Windows workstation. > > Granted the comparison might not quite be apples vs apples in a true > scientific sense, at least it would be slightly more scientific than > some of the anecdotal information shared by some here. > > so perhaps green apples vs red apples :-) > > > 2009/5/13 Digy <[email protected]>: >> I think, at first, it is just a simple comparison of results of the same >> code with the same index on the same hardware. >> >> After then, some optimizations specific to to .Net or Java (or more complex >> test cases) can be thought. >> >> DIGY. >> >> >> -----Original Message----- >> From: Todd Carrico [mailto:[email protected]] >> Sent: Wednesday, May 13, 2009 10:57 PM >> To: [email protected] >> Subject: RE: Lucene Java Vs .Net >> >> My issue with this type of comparison is the difference in hardware. Small >> differences can make a big impact. >> >> What are you doing to make sure this is apples to apples as far as hardware >> goes? >> >> tc >> >> -----Original Message----- >> From: Digy [mailto:[email protected]] >> Sent: Wednesday, May 13, 2009 2:48 PM >> To: [email protected] >> Subject: RE: Lucene Java Vs .Net >> >> Hi Shashi >> >> [email protected] could be a good place to discuss your tests and >> results. >> I eagerly wait your results. >> >> DIGY >> >> >> -----Original Message----- >> From: Shashi Kant [mailto:[email protected]] >> Sent: Wednesday, May 13, 2009 3:21 AM >> To: [email protected] >> Subject: Re: Lucene Java Vs .Net >> >> Hi, I am bumping up this thread to see if there are any further inputs >> on this. I was planning on running a benchmarking test for .NET vs >> Java for both Indexing & Querying (single field, multi field, span >> etc.), and would be happy to share my results. I would be using the >> PubMed corpus of biomedical literature and running indexing on a few >> million articles. >> >> I would be interested to see if others are running on other corpora, >> sharing ideas would be very helpful IMHO. >> >> Regards, >> Shashi >> >> >> >> >> >> On Fri, May 8, 2009 at 12:54 PM, Matt Honeycutt <[email protected]> >> wrote: >>> I haven't compared Lucene to Lucene.NET, but I have seen huge performance >>> improvements in Weka on .NET vs. Java. Taking the Java jar files and >> using >>> IKVM.NET to produce a .NET assembly resulted in a huge performance >>> improvement. I would fully expect Lucene.NET to hold a similar >> performance >>> advantage over its Java equivalent. >>> >>> On Fri, May 8, 2009 at 11:42 AM, Granroth, Neal V. < >>> [email protected]> wrote: >>> >>>> I'd also be interested in hearing about performance differences between >> the >>>> two. Out of idle curiosity I run a few very basic trials. I expected >> there >>>> to be no noticeable difference, but was surprised to find the .NET >>>> application ran 3x faster than the Java equivalent. >>>> >>>> -- Neal >>>> >>>> -----Original Message----- >>>> From: [email protected] [mailto:[email protected]] On >>>> Behalf Of Wayne Douglas >>>> Sent: Tuesday, May 05, 2009 3:46 AM >>>> To: [email protected]; [email protected] >>>> Subject: Re: Lucene Java Vs .Net >>>> >>>> Does anyone have any benchmark data on the performance of these two - >>>> purely >>>> out of interest :) >>>> >>>> On Mon, May 4, 2009 at 10:26 PM, Andreas Mummenhoff <[email protected]> >>>> wrote: >>>> >>>> > yes, >>>> > Lucene.Net is a class-per-class, API-per-API and algorithmatic port of >>>> the >>>> > Java Version. The release numbers are the same, so Java Lucene 2.3.1 ~= >>>> > Lucene.Net 2.3.1 >>>> > And the consequence of this is index compatibility, so you can search >> and >>>> > fill a Lucene.Net index with Java Lucene and the other way round. >>>> > >>>> > Andreas >>>> > >>>> > -----Ursprüngliche Nachricht----- >>>> > Von: MBVV.Satish [mailto:[email protected]] >>>> > Gesendet: Montag, 4. Mai 2009 22:47 >>>> > An: [email protected] >>>> > Betreff: Lucene Java Vs .Net >>>> > >>>> > >>>> > >>>> > Hi, >>>> > We are currently using Java based Lucene in one of our products. We >> have >>>> a >>>> > product in Microsoft .Net where we intend to use Lucene.Net. >>>> > My question is.... >>>> > 1.) Is Lucene.Net has all the features of Java version? or is there any >>>> > specific limitation to .Net version? >>>> > 2.) Is Lucene.Net is class-per-class, API-per-API and algorithmatic >> port >>>> > of Java Version? if so are the releases comparable by release numbers >>>> i.e., >>>> > .Net 2.3 with Java 2.3? >>>> > 3.) Is it possible to use the same Lucene indexes in Java Version as >> well >>>> > as .Net Version? >>>> > >>>> > Please answer my above questions so that it will be helpful in >> deciding >>>> on >>>> > the use of Lucene .Net. >>>> > >>>> > Thanks >>>> > Satish >>>> > >>>> > >>>> > Now surf faster and smarter ! Check out the new Firefox 3 - Yahoo! >>>> > Edition http://downloads.yahoo.com/in/firefox/?fr=om_email_firefox >>>> > >>>> > >>>> >>>> >>>> -- >>>> Cheers, >>>> >>>> w:// >>>> >>> >> >> >
