WHERE IN performance

[Knepley, Jim]

Is it anyone elses experience that queries with large IN stanzas in a
WHERE clause don't scale very well?
It seems like it's beyond a linear performance hit when I have a large
number (thousands) of tokens in an IN clause, even when the matching
field is indexed.
 
Is this something that buffer tweaks can mitigate, or am I looking at a
more fundamental issue with how IN in implemented?
 
J