On Thu, Mar 17, 2011 at 9:13 AM, Jeff <threshar@torgo.978.org> wrote:
> hey folks,
>
> Running into some odd performance issues between a few of our db boxes.
> While trying to speed up a query I ran it on another box and it was twice
> as fast. The plans are identical and various portions of the query run in
> the same amount of time - it all boils down to most of the time being spent
> in a join filter. The plan is as good as it is going to get but the thing
> that is concerning me, which hopefully some folks here may have some insight
> on, is the very large difference in runtime.
My experience puts the 23xx series opterons in a same general
neighborhood as the E5300 and a little behind the E5400 series Xeons.
OTOH, the newer Magny Cours Opterons stomp both of those into the
ground.
Do any of those machines have zone.reclaim.mode = 1 ???
i.e.:
sysctl -a|grep zone.reclaim
vm.zone_reclaim_mode = 0
I had a machine that had just high enough interzone communications
cost to get it turned on by default and it slowed it right to a crawl
under pgsql.