hm, during new runs we see only 2x slow down. So, no "much slower",
but just "2x slower" -- both searches compared are done in single
thread.
best regards
--
Valery A.Khamenya
On Sun, Mar 28, 2010 at 2:09 PM, Valery Khamenya <khame...@gmail.com> wrote:
> yes, the same parameters.
>
> best regards
> --
> Valery A.Khamenya
>
>
>
> On Sun, Mar 28, 2010 at 12:18 AM, Andi Vajda <va...@apache.org> wrote:
>>
>> On Sat, 27 Mar 2010, Valery Khamenya wrote:
>>
>>> there is a strange slowdown during the search for similar documents.
>>> For some reason pylucene version is much slower than the pure Lucene one.
>>> The test document collection contains 200K docs.
>>>
>>> Here is the pylucene version:
>>>
>>> content = ref_doc.getField('content').stringValue()
>>> similarity_query = SimilarityQueries.formSimilarQuery(content,
>>> default_analyzer, 'content', None)
>>> search = index.search(similarity_query, 200)
>>
>> Did you initialize the JVMs with the same memory parameters when comparing
>> PyLucene vs Lucene Java ?
>>
>> Andi..
>>
>
