On Thu, May 17, 2012 at 11:35 AM, Joshua Berkus <[email protected]> wrote: > Jeff, > > That's in-RAM speed ... I ran the query twice to make sure the index was > cached, and it didn't get any better. And I meant 5X per byte rather than 5X > per tuple.
Ah, OK that makes more sense. I played around with this, specifically count(*), quite a bit when IOS first came out, and I attributed a large part of the time to the code that forms a tuple out of raw bytes, and the code that advances the aggregate. The first one is probably more a per-tuple cost than per byte, and the second definitely is per tuple cost. I can't find my detailed notes from this work, so this is just from memory. Cheers, Jeff -- Sent via pgsql-hackers mailing list ([email protected]) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
