Re: Splitting queries across servers

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Max" <maxdl(at)adelphia(dot)net>
Cc: pgsql-general(at)postgresql(dot)org
Subject: Re: Splitting queries across servers
Date: 2005-01-29 23:11:58
Message-ID: 27480.1107040318@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

"Max" <maxdl(at)adelphia(dot)net> writes:
> Here's a question for the developers: what's the memory consumption
> difference when you move a 3GB database from a 32 bit machine to a 64 ?

Not a lot.

> Isn't the whole data taking up more RAM because pointers are now 64 bits
> instead of 32 ?

There are no pointers in on-disk data, so there's no difference at all
in file sizes, and thus none in shared-buffer requirements --- at least
not from the size of pointers. (64-bit machines tend to have MAXALIGN
of 8 bytes instead of 4, which implies more alignment padding at the
ends of rows. But that's usually a few-percent kind of cost, not a
factor of 2.)

The pointer size difference has a larger influence on Postgres' other
internal data structures, such as the shared lock table. But by
comparison to the shared disk buffers, those generally don't amount
to anything.

regards, tom lane

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Tzahi Fadida 2005-01-30 00:17:12 Pinning buffers for long times like outer joins might do.
Previous Message Bruno Wolff III 2005-01-29 22:41:22 Re: inet-type sequence