On Wed, May 11, 2011 at 4:38 PM, Merlin Moncure <mmoncure@gmail.com> wrote:
> Following are results that are fairly typical of the benefits you
> might see when the optimization kicks in. The attached benchmark just
> [hbcache]
> real 3m35.549s
> [HEAD]
> real 4m24.216s
These numbers look very good. Thanks for responding to my request.
What people have said historically at this point is "ah, but you've
just deferred the pain from clog lookups".
The best way to show this does what we hope is to run a normal-ish
OLTP access to the table that would normally thrash the clog and show
no ill effects there either. Run that immediately after the above
tests so that the cache and hint bits are both primed.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
