On Tue, Jun 15, 2010 at 8:23 AM, Matthew Wakeling <matthew@flymine.org> wrote:
> Absolutely, and I said in
> http://archives.postgresql.org/pgsql-performance/2010-03/msg00272.php
> but applied to the Postgres B-tree indexes instead of heaps.
This is an interesting idea. I would guess that you could simulate
this to some degree by compiling PG with a larger block size. Have
you tried this to see whether/how much/for what kind of workloads it
helps?
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company