[postgis-devel] Forget what I said

[Paragon Corporation]

Its slower

SELECT count(p1.*) 
from par35 As p1 INNER JOIN par35 As p2 ON p1.the_geom && p2.the_geom
WHERE p2.map_id LIKE '0101%';

On gserialized: 1,076 ms

On regular (9.1 though no gserialized) -- 110ms

So I guess the network effects of returning the dataset was taking up the
chunk of my last timing.

I still need to do a 9.0 to 9.0 compare to be fair.