On Mon, 13 Jul 2009, Suvankar Roy wrote:
> I believe that execution time in greenplum should be less compared to postgres.
Well, first off you don't even mention which PostgreSQL or Greenplum
version you're comparing, which leaves a lot of variables we can't account
for. Second, you'd need to make sure that the two servers had as close to
identical server parameter configurations as possible to get a fair
comparison (the postgresql.conf file). Next, you need to make sure the
data has been loaded and analyzed similarly on the two--try using "VACUUM
ANALYZE" on both systems before running your query, then "EXPLAIN ANALYZE"
on both setups to get an idea if they're using the same plan to pull data
from the disk, you may discover there's a radical different there.
...and even if you did all that, this still wouldn't be the right place to
ask about Greenplum's database product. You'll end up with everyone mad
at you. Nobody likes have benchmarks that show their product in a bad
light published, particularly if they aren't completely fair. And this
list is dedicated to talking about the open-source PostgreSQL versions.
Your question would be more appropriate to throw in Greenplum's direction.
The list I gave above is by no means even comprehensive--there are plenty
of other ways you can end up doing an unfair comparison here (using
different paritions on the same disk which usually end up with different
speeds comes to mind).
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD
