Re: 8.3devel slower than 8.2 under read-only load

On Wed, 21 Nov 2007, Guillaume Smet wrote:

> *** 8.2 ***
> tps = 853.360277 (including connections establishing)
>
> *** 8.3 ***
> tps = 784.819087 (including connections establishing)

This is an 8% drop. I've seen a larger difference than that between two
identical installations of the same version when the database is many GB
large. Hard drives deliver a higher transfer rate at their inner
portions, typically the start of the disk from the operating system's
perspective. It's not unusual for the slow parts of the disk to be 30-40%
slower than the fast ones. I've been known to mkfs all the database
paritions before each test run just to remove this bias, so that the data
was on exactly the same portion of the drive each time.

Not saying this is responsible for your results, just that benchmarking is
hard and there may be somthing other than what you think responsible for a
difference of this size. I'd suggest running "select count(*) from x" on
a couple of the big tables as one way to get a feel for whether the
underlying disk is delivering at the same speed in both installations.

--
* Greg Smith gsmith(at)gregsmith(dot)com http://www.gregsmith.com Baltimore, MD