tsearch2 memory problem

Hi!

I realized that loading a dictionary with ~160000 words consumes
additional ~40 MB of memory for each connection. It obviously doesn't
use a shared memory. Is it possible to decrease the memory consumption?
I found this thread

http://www.mail-archive.com/pgsql-general@postgresql.org/msg116924.html

but no solution. Is there any progress in this area?

Thanks.

David