Did you try to insert the xml data into a table with an indexed xml column?
AFAIK, SQL Server does not calculate any statistical information on a variable. Therefore, it can't detect that the pre-defined(?) value is off.
By storing the data in an indexed xml column, you might benefit from the statistic information.
As a side note: I don't think this will lead to parallelism though, since 15k elements is still not that much.
Most probably, storing the data in an indexed xml column will already improve performance so parallelism is no longer needed.
If my guess turns out to be wrong, please provide more detailed information (sample data, current query, expected result, current performance information) so we can have a closer look into it.
A pessimist is an optimist with experience.
How to get fast answers to your question[/url]
How to post performance related questions[/url]
Links for Tally Table [/url] , Cross Tabs [/url] and Dynamic Cross Tabs [/url], Delimited Split Function[/url]